trend is that the wandisco plc october 19, 2017 · wandisco cmd; veni, vidi, cc decies centena cmd...

Completed: 19 October 2017 01:47EDTDisseminated: 19 October 2017 01:47EDT

George O'Connor | +44 (0) 20 7710 7694 | george.o'[email protected] Sales Desk | +44 (0) 20 7710 7600

MiFID II - Research: Is your access agreed? CONTACT us today

October 19, 2017 European Software & IT Services

WANdisco plcWAND – LSE; WAND.L

BUY

COMPANY UPDATE WANdisco CMD; Veni, vidi, CC decies centenaPrice (18 October 2017) 863p

Changes

Rating

Target Price

Previous

-

-

Current

BUY

1,029p

Key data

Bloomberg/Reuters codes: WAND LN / WAND.L

Market cap (£m) 287

FTSE ALL SHARE 4,141

1mth perf (%) 29.2

3mths perf (%) 37.5

12mths perf (%) 456.5

12mth high-low (p) 890 - 128

Free float (%) 71

Key financials

Year to Dec 2016A 2017E 2018ESales 11.4 16.8 20.1EBITDA (adj) (7.5) (2.4) 0.0EV/EBITDA (x) 0.5 1.8 (171.9)PE adj (x) NA NA NAEV/Sales (x) (0.3) (0.3) (0.1)EPS adj (c) (0.47) (0.25) (0.20)

Prices are as of close 18 October 2017

All sources unless otherwise stated: Companydata, FactSet, Stifel estimates

Share price performance (indexed)

700

600

500

400

300

200

100

0Nov-16 Jan-17 Mar-17 May-17 Jul-17 Sep-17

Absolute Rel.FTSE ALL SHARE (ASX)

SummaryLike Julius Caesar conquering Pharnaces II, Tuesday’s WANdisco CMD was a swift,conclusive affair. IBM fired the killer blow by suggesting that it could see US$200-300mrevenue from its WANdisco OEM. For us, other highlights include: (i) the convincingnarrative that WANdisco has the capabilities to execute on the opportunity. (ii)Commentary that “the puck” is moving to WANdisco not the other way around. (iii) Fromthe technical presentations more clarity around the defensibility of the technology, but alsovery pleasing to see WANdisco open the kimono on its longer range thinking. (iv) From theesteemed industry analyst Peter Burris there was an overlay message that the industrytrend is that the “cloud goes to where the data” – which we read as an endorsement ofthe importance of hybrid computing (remember for some it is only a transitionary step –we think that it is ‘the thing’). On the grudge list, no customer presentation, no colour onQ3 trading, nothing on SCM (a positive from H1 results) and no commentary on the deepchanges wrought by CEO (those ‘99 Problems’). This has been a masterful turnaroundand its fixing process deserves more mention. As we return to our desk we have revisited(i) our DCF to try to accommodate some of the IBM ‘what it is’ thinking and, (ii) giventhe Microsoft Data Box news on Tuesday the evidence says that WANdisco has a criticalpiece of industry plumbing, is an asset to a collection of deep pocket behemoths andso we revisited our SOTP. Net/net these changes increase our target price from 843p to1029p. We retain our Buy rating.

Key PointsForecasts. We make no changes to our forecasts or recommendation.

IBM. Mr. Dimtchev spoke to previous cases where IBM had taken a US$20-30m revenuebusiness and transformed it into a US$200-300m in a few years and that was the intentionwith IBM Big Replicate (i.e. the WANdisco Fusion OEM product).

Industry analyst Mr. Burris pointed out that total cost of ownership studies had shownthat for some the cost of data transfer is currently prohibitive, and as processing datalocally is more cost effective, it is the cloud that comes to the data, rather than the currentorthodoxy which is that the data goes to the cloud.

Ability to execute. While the ‘puck moves to WANdisco’ helps TAM and makes the usecase more tangible, our long standing grudge is that it is not TAM but rather the abilityto execute on TAM which is the more critical component in judging success. WANdisconeeds to be able not just to onboard customers, but also help on sales prospects andPOCs from the channel prospects and POCs at scale and at speed. On this, we believeTuesday’s CMD was a strong positive.

Impressive technology – read all about it. This is all possible because of WANdisco’s IPstrength with DConE, its patent-protected ‘active transaction replication’ software (allowsdata to be moved securely, at speed and at scale between computing environments).WANdisco Fusion is a general purpose replication platform with plug-ins for multiplestorage systems (cloud, big data, etc) rather than a specific platform for Hadoop.

Watch again. In case you missed the CMD it is here. http://webcasting.brrmedia.co.uk/broadcast/59d65608d349960788385d17/59e6fee232342194530000c5

Target price update 1029p. WANdisco offers investors exposure to a business enjoyingrigorous growth in a global target market, and to perhaps the ‘noisiest’ theme in ITcurrently – Big Data. We see abundant evidence that the company is now positioned toaccelerate growth that should create further value for its shareholders. We use a blendedmodel to arrive at our 12-month share price target of US1,092cents/843p (taking DCFUS1526cents (increased from US1,145cents), sum-of-the-parts 1597cents (increasedfrom US1,110cents) and FCF yield US1,251cents).

All relevant disclosures and certifications appear on pages 21 - 23 of this report.

Stifel does and seeks to do business with companies covered in its research reports. As a result, investors should be awarethat the firm may have a conflict of interest that could affect the objectivity of this report. Investors should consider thisreport as only a single factor in making their investment decision.

mailto:[email protected]?subject=MiFID II research: Is your access agreed?&body=I would like to be contacted regarding continuing access to Stifel research after implementation of MiFID II.

http://webcasting.brrmedia.co.uk/broadcast/59d65608d349960788385d17/59e6fee232342194530000c5

http://webcasting.brrmedia.co.uk/broadcast/59d65608d349960788385d17/59e6fee232342194530000c5

Key data1

Key profit and loss data ($)

2016A 2017E 2018ESales 11.4 16.8 20.1EBITDA (9.3) (3.7) (1.4)EBITDA margin (%) (81.3) (21.8) (6.8)Gross profit 10 16 19Net income (9.2) (12.9) (8.6)PBT rep (10.0) (13.8) (9.6)EBITDA adj (7.5) (2.4) 0.0Depreciation & amortisation 9 8 8DPS 0.00 0.00 0.00

Key cash flow data ($)

2016A 2017E 2018EOperating profit (17.9) (11.5) (9.6)Operating cash flow (2.2) 4.7 3.7Capex (0.1) (0.1) (0.1)Dividends 0.00 0.00 0.00Net debt (7.6) (8.2) (6.8)Taxes paid (0) 1 1Cash flow from investing 5 1 (1)FCF (8.3) 0.6 (1.4)Dividends 0.00 0.00 0.00Cash at end of year 8 8 7

Key balance sheet ($)

2016A 2017E 2018ECash and cash equivalents 7.6 8.2 6.8Total assets 20.2 20.2 18.5

Key informationTarget price methodology/risks

We use a blended model to arrive at our 12-month share price target of1029p, using discounted cash flow, sum-of-the-parts and free cash flowyield. While WANdisco has a number of adjacent growth opportunities,we believe the ‘cash generation’ bias in our valuation methodologyreflects how investors see the benefits of subscription-based businessmodels.

Risks to target price. In addition to general and macroeconomic risks,the downside risks include continued deceleration in the source codeand Big Data markets. This would impact cash inflow and therebyincrease cash outflow and lessen investor interest. Upside risks includea better-than-expected revenue growth, possibly as a consequence ofthe channel partner sales accelerating faster than anticipated.

Business description

WANdisco is an infrastructure software company that has developed apatent-protected method for data replication of across heterogeneouscompute environments.

Senior management

David Richards - CEO

Erik Miller - CFO

Key dates

Jan 2018 - Trading update

Major shareholders

Oppenheimer Funds 15.02%

Schroder Investment Management 9.86%

T. Rowe Rice International 6.09%

GAM 4.38%

Ross Creek Capital 3.88%

Website

www.wandisco.com1 Year end December

Data in millions, except per share and percentagesSource: Company data, FactSet, Stifel estimates

WANdisco plcWAND – LSE

European Software & IT ServicesCompany Update

October 19, 2017

2

WANdisco CMD; Veni, vidi, CC decies centena

CMD – In summary Peter Burris, GM Wikibon.

Mr Burris commented that for digital businesses, ‘data’ is the basis of their differentiation and for them,

data is an asset and their objective must be to put that asset to work and apply it to new areas. Hence

their need to unlock data from multiple places. In addition, as data is not scarce, nor is it fungible, it has

a different value when used in different applications, and so by its nature it will, over time, be more

distributed. As digital businesses create new higher value applications, their data will be increasingly

more distributed (part of the rationale being that costs are lower) and hence there is a growing need to;

(i) be able to restore data and, (ii) protect it. According to Mr Burris this is an under-served part of the

market.

While the cloud is a critical architectural component for digital business, Mr Burris argued that not all

data will move to the cloud and pointed to total cost of ownership studies which concluded that for some

the cost of data transfer is currently prohibitive, and also processing data locally is more cost effective.

Hence his view that “the cloud comes to the data” – which goes against the current orthodoxy that the

data goes to the cloud.

Mr Burris argued that over time the IaaS market (say Amazon AWS) will not out-grow the SaaS or

private cloud markets (see figure below) because the cloud moves to the data – again Mr Burris broke

with the current orthodoxy. As data will drive greater functional distribution this will lead in turn to greater

application complexity (see figure), more distributed data and this will increase the importance of; (i)

protecting data, (ii) near real-time data replication, (iii) application-level data restoring, and (iv) the

importance of data governance – where users will want to ensure that the data is not ‘locked-in’ by any

particular IaaS/PaaS vendor. This market opportunity would be worth “at least US$1bn”.

Looking at Disaster recovery, Mr Burris argued that his research concluded that on average applications

fail to restore c35% of the time. For a typical Fortune 100 firm this results in US$3.7bn lost revenue and

for a digital business failure to restore means that the business is lost. This market is judged to be

cUS$8bn market and one characterised by lots of customer churn.

Figure 1: Worldwide Enterprise IT Projection by vendor revenue segmentation, 2015-2026 (US$bn)

Source: Wikibon, 2017



October 19, 2017

3

Figure 2: Illustrating application complexity

Source: Wikibon, 2017

Jagane Sundar, CTO, WANdisco

Mr Sundar opened the kimono as to WANdisco's longer run (i.e. two years plus) thinking regarding the

product direction. By definition this is outside the usual forecast period, but amongst the commentary

was; (i) data has expanded beyond single servers to clusters and now demand is pushing for data to be

available across multiple regions (this dovetailed with Mr Burris points on data distribution) and that

without WANdisco demanding consistent coherent, always available distributed data applications is

impossible.

Regarding the longer run product roadmap Mr Sundar talked through the importance of; (i) delivering

higher throughput rates, (ii) applying Paxos to database transactions and (iii) ‘pipelining’ or reducing

latency on data transfer.

Paul Scott-Murphy, VP Product Management, WANdisco

Mr Scott-Murphy begin with a nice recap to remind all that WANdisco Fusion is a general purpose

replication platform with plug-ins for multiple storage systems rather than being designed to be specific

to Hadoop (investors, we find, still see WANdisco as a ‘Hadoop play’), which enables replication of

changing data to the cloud and on-prem.

Mr Scott-Murphy highlighted the four WANdisco Fusion use cases as: (i) data centre to data centre, (ii)

data centre to cloud, (iii) physical data movement – note the Amazon Snowmobile and Microsoft Data

Box announcements and, (iv) Cloud to cloud. For us this linked with Mr Burris’ point on the importance

that users are starting to attach to moving data from IaaS to IaaS vendor (say from AWS to Azure) in

order to avoid and vendor lock-in or predatory pricing. [We commented this week that as IaaS services

were all starting to look the same, for users they should be thinking about exploiting price deflation –

here https://stifel2.bluematrix.com/sellside/EmailDocViewer?encrypt=75b28fbc-0bbb-4992-afcd-

fcd1805264a0&mime=pdf&co=Stifel&id=george.o'[email protected]&source=mail

As to the shorter-term product roadmap we were encouraged by the moves to engage the wider

software developer community be that in enabling better application independence, OEM readiness and

a development kit (in industry jargon STK) which will enable more third party development. For us this

would be a very significant departure.

Keith Graham, SVP Worldwide Sales WANdisco

Mr Graham talked through the switch at WANdisco from a direct sales organisation to developing an

indirect sales model and how this resulted in a different sales process. We noted positively a very

flexible partner model which seemed to include any version of partner model (spanning referral, reseller,

fulfilment, marketplace, co-selling and OEM). On the sales execution side we noted that WANdisco’s

commitment was deeper than we had expected with; (i) Lead generation in EMEA and North America,

(ii) Sales and pre-sales in North America, EMEA and APAC, and (iii) channel Manager for IBM

Americas, IBM EMEA, Oracle WW and AWS. [Aside: we spoke to IBM about this afterwards and they

were very happy that WANdisco had hired ex-IBMers to manage the IBM relationship.]



October 19, 2017

4

https://stifel2.bluematrix.com/sellside/EmailDocViewer?encrypt=75b28fbc-0bbb-4992-afcd-fcd1805264a0&mime=pdf&co=Stifel&id=george.o'[email protected]&source=mail

https://stifel2.bluematrix.com/sellside/EmailDocViewer?encrypt=75b28fbc-0bbb-4992-afcd-fcd1805264a0&mime=pdf&co=Stifel&id=george.o'[email protected]&source=mail

Mr Graham concluded with a case study of the sales model whereby ‘home team’ had been working

with a particular client on a 100 node cluster, yet they went with IBM who was working with the same

client as part of a larger deal. The eventual contract was expanded to 750 nodes.

Peter Scott, SVP Business Development, WANdisco

Mr Scott is one of the veterans at WANdisco (c10 years) and he is accredited with driving the pivot to an

indirect sales model in 2015. Mr Scott gave colour on the underlying ‘why’ thinking in the sales model

and in particular talked though the importance of the strategic partners, such as AWS, IBM, Microsoft,

Oracle. How each had a different engagement model and how WANdisco has a multi-stage model for

on-boarding new channel partners. He also commented on the importance of Alibaba and growing data

volumes in that region. Alibaba is not a current partner. [As a quick illustration as to likely data volumes:

there are 751m Internet users in China, 462m in India and 286m in US (source Internetworldstats.com)].

Regarding IBM, Mr Scott reminded that the success of the partnership was illustrated by the H1/2017A

US$3.5m royalty and that the pair had a significant pipeline of POCs (in industry jargon: proof of

concept – usually a step before a customer buys a license) for H2/2017 with IBM. [This point was re-

iterated by IBM in their presentation.]

Mr Scott concluded that WANdisco had a deep competitive moat, and a strong channel partner

ecosystem.

Nick Dimtchev, Business Unit Executive, IBM/Hortonworks Alliance Global Sales Leader

The IBM/Hortonworks Alliance group is part of IBM’s analytics division. Mr Dimtchev spoke to previous

cases where IBM had taken a US$20-30m revenue business and transformed it to a US$200 - 300m

one in a few years. That was the intention with IBM Big Replicate (i.e. the WANdisco Fusion OEM

product). Mr Dimtchev talked how IBM has made this a repeatable structured process which starts with

the customer use case and how much they will pay. Dovetailing with Mr Burris earlier comments, Mr

Dimtchev stated that distributed data, different types of data and ensuring that the customer has the

right data to run their applications is a critical success factor. Mr Dimtchev stated that the opportunity

was not about implementing dashboard technology but rather about managing data and data

governance. Of the comments:

The importance of the ‘data landing zone’. IBM has signed two OEM agreements in this

business area, WANdisco, and more recently Hortonworks. IBM sees Hortonworks as the number

1 Hadoop distribution.

Why did IBM chose WANdisco? (i) It supports multiple databases and so gives IBM the flexibility

to be heterogeneous, like the target customer base. (ii) There are no other enterprise-grade

alternatives for restore and data replication. (iii) WANdisco Fusion solves big enterprise user pain

points and it has a tangible ROI.

Who are the target customers? These are typically regulated industries such as Banking,

Insurance, Telco and Manufacturing, where the customers are complex and large (20 – 4,000

Hadoop nodes) and in production mode (i.e. they have moved from using Hadoop for test and

development). They are all looking to reduce cost of ownership and risk.

Why does IBM BigReplicate win? (i) It is not just one product and customers like the ‘one-stop

shop’ integrated product as part of the wider IBM product family, (ii) it plays a role as part of IBM’s

hybrid cloud strategy – thereby helps customers migrate at their pace, (iii) the IBM wider focus and

investment in big data, i.e. analytics and Watson (iv) the strategic alliance with Hortonworks.

The customer use case. Mr Dimtchev talked though a use case based on downtime which had

been costing the customer US$466k per incident. For this customer over three years IBM was able

to demonstrate savings of US$9.3m.



October 19, 2017

5

How many potential customers? Including the c100 Hortonworks customers, Mr Dimtchev talked

about up to 2,000 customers and that IBM were looking to extend outside North America and

Europe. The deals would be very large transactions, i.e. multi-million dollar value deals. In Q&A,

WANdisco CEO David Richards talked about a US$700m deal (value to IBM contract) and later

chatting with Mr Dimtchev he also talked though other similarly very large deals.

Figure 3: IBM Value adds for Big Data

Source: IBM, Stifel Research

WANdisco Fusion is here



October 19, 2017

6

Stifel Viewpoint

This is part of a continuum…

We are reminded again of that well-used phrase, “Skate to where the puck is going, not where it has

been” as we reflect on Tuesday’s CMD. In essence, we believe that this reflects the dramatically

increasing (i) interest in data migration to the hybrid cloud, and (ii) the new importance of high-values

services, such as Disaster Recovery and high availability and data migration technologies amongst the

channel partner companies (i.e. Amazon, Microsoft, IBM, Google) and large end user organisations.

Furthermore, the news builds on a rack of positive news from WANdisco in the past couple of months,

where we have charted its expanding customer sales progress and the competitive USP

…which tells us about the growing importance of hybrid cloud…

The ‘puck’ is heading to hybrid cloud and the marketplace driver is the growing acceptance of hybrid

cloud and use of IaaS has transformed how users procure and manage compute and storage

infrastructure. This is in turn leading to increased investments and greater competition among a number

of public cloud IaaS providers and heightened strategic focus on their offerings – and this will continue

to build out channel interest in working with WANdisco. While IaaS services are starting to look similar

across the top providers, the strengths of the ecosystem of tools will be a strong way in which public

cloud IaaS providers differentiate and position themselves in this market – hence the importance of

ensuring that they have an ‘active replication solution’.

And the change in storage formats (NAS &SAN give way to Object storage in a cloud setting) all help

We also note a trigger in the rise of object storage, which has become prominent in the form of array-

type products as well as being the basis for cloud-based protocols such as Amazon’s S3. To see how

object storage differs significantly from SAN and NAS protocols, File and block are file system-based

methods of storage access. In both cases, there is a file system which organise data into files and

folders in a tree-like hierarchy. But under the bonnet, that file path and the file system also handle

addressing to the physical location of blocks of storage on the media itself. The key difference between

file access/NAS and block access/SAN is that in NAS, the file system resides on the array. In a SAN,

the file system is external to the array and I/O calls are handled by the file system on the server, with

only block-level information required to access data from the SAN. To briefly compare and contrast:

NAS: Good at secure file sharing. Can become siloed. Scale-out NAS potentially good at scale.

Bad at extreme scale.

SAN: Good at transactional and database workloads. Can be expensive.

SAN and NAS: Both can come with advanced storage features, such as replication. Both can be

relatively costly compared with object storage on commodity hardware, although both SAN and

NAS software-defined storage are available. Both lack the rich metadata of object storage.

Object storage: Very scalable, suited to unstructured data and large datasets, potentially good for

analytics via rich metadata. Lacks high-end performance and data protection is slow across

clusters. Can be very cost-efficient, hardware-wise.

Sales execution is critical

WANdisco is a products company. Its pivot from direct to indirect sales has been accomplished without

it losing that 'product centricity' - everything comes from that and it is critical to its long term success.

While the ‘puck moves to WANdisco’ helps TAM and makes the use case more tangible, our long

standing grudge is that it is not TAM but rather the ability to execute on TAM which is the more critical

component in judging success. WANdisco needs to be able not just onboard customers, but also help

on sales prospects and POCs coming through from the channel prospects and POCs at scale and at

speed. On this, we believe Tuesday’s CMD was a strong positive.



October 19, 2017

7

INVESTMENT CASE – revisited

WANdisco is an infrastructure software company operating at the confluence of four IT axes: big/lots of

data; migration to Cloud; importance of RASS; and agile software development/DevOps. The company

has differentiated technology and an impressive client list, and is currently enjoying strong operational

momentum.

Impressive technology

WANdisco replicates data across heterogeneous environments. Its Distributed Coordination Engine

(DConE) is patent-protected (11 issued, 27 pending). DConE is capable of active transaction replication,

where data servers are equal peers in a distributed network. This means data is never lost (critical in

disaster recovery situations), and can be moved, at speed and at scale, between computing

environments (critical in replication and data migration scenarios) and in new-world Cloud SLA

management (critical for customers migrating between IaaS providers). Customers talk about a very

strong ROI. This technology is an enhancement of the Paxos algorithm to enable active-active

replication between a variety of data sources including Hadoop clusters, Cloud environments, NAS

(network-attached storage) filers, etc. It enables continuous data access in the face of network outages,

hardware failures and entire data centres going up and down.

The bluest of the blue-chip customers

WANdisco’s customers are companies that are building software, are organised globally, and are

migrating to the Cloud – this should make for a very large total addressable market (TAM). Sectors

include: auto, entertainment, financial services, government, healthcare, IT, telecoms and utilities. The

customer list has a strong tech constituency that includes not only Accenture, ARM, Cisco, Dell, HP and

IBM, but also banks like HSBC and such global brands as GE, Fidelity, John Deere, Johnson &

Johnson, Pitney Bowes and Wal-Mart. The company has c200 customers (of which c31 are on

WANdisco’s Big Data product, and most use its ALM product).

Large TAM still expanding organically

Trying to draw a boundary around WANdisco’s TAM can be a frustrating exercise. Thinking through the

technology tends to throw up new use cases in multiple adjacent customer markets and vertical

industries, all in addition to the current focus on replication, migration and Application Lifecycle

Management. Looking at the three core areas, we think: (1) disaster recovery should be a cUS$11bn

TAM; (2) data migration to the Cloud suggests a cUS$7bn TAM; (3) inter-Cloud data replication,

availability and procurement/SLA management should be a large and viable market as the cloud-based

application hosting market matures and enterprise users think about their pricing power.

Attractive multiples – all about the growth

WANdisco is enjoying accelerating growth. The H1 numbers headline with news of record bookings

+97% YoY, 109% growth in H2 bookings, FY total bookings +72% YoY. Following a US$14m fund

raising in summer 2016, cash stood at US$7.6m on 31 December, up from US$1.1m at 30 June 2016.

The new multi-layered sales and distribution model is thriving

A key focus for 2016 was to establish a partner network. Mission accomplished – now, in addition to its

own ‘direct sales team’, WANdisco has a set of Tier 1 channel partners. This includes IBM, with which

WANdisco inked a rare OEM agreement, as well as significant channel partnerships with Oracle and

Amazon. All are contributing to bookings and are instrumental in building the company’s sales pipeline.

They also reduce the cost base, thereby hastening WANdisco along the path to profitability.

Management stays the course, maintains the ‘passion’

Management has been through the mill – and remains together. The ‘top table’ still includes founders

David Richards and Yeturu Aahlad. New CFO Erik Miller has public and private software industry

experience.



October 19, 2017

8

Figure 4: WANdisco: Key data (US$m)

Year to 31 December 2015A 2016A 2017E 2018E 2019E 2020E

Bookings 9.00 15.50 20.83 29.29 40.60 50.35

Sales 10.99 11.38 16.81 20.10 24.29 29.11

Gross profit 10.25 10.03 15.64 18.69 22.59 27.07

EBITDA -15.99 -7.46 -2.41 0.02 2.84 6.74

EBIT -29.92 -17.89 -11.49 -9.58 -7.32 -4.01

Continuing pre-tax profit -16.76 -7.91 -2.62 -0.20 2.60 6.47

Reported pre-tax profit -31.04 -10.05 -13.78 -9.58 -7.32 -4.01

Tax rate 0.07 0.10 0.31 4.20 -0.34 -0.14

Net Income -29.91 -9.27 -12.97 -8.73 -6.42 -3.08

Net funds 2.56 7.26 4.89 3.42 3.88 9.32

Per Share Data (c)

Fully diluted adjusted EPS -0.88 -0.47 -0.25 -0.20 -0.13 -0.04

DPS 0.00 0.00 0.00 0.00 0.00 0.00

Free Cash Flow -90.80 -25.07 1.70 -3.82 1.27 5.53

Free cash flow yield -7.9% -2.2% 0.1% -0.3% 0.1% 0.5%

Valuation Data

EV/Bookings 48.87 28.08 21.00 14.99 10.80 8.60

EV/Sales (x) 40.01 38.24 26.02 21.84 18.05 14.88

EV/EBITDA (x) -27.51 -58.30 -181.16 25461.80 154.39 64.27

EV/EBIT (x) -14.70 -24.32 -38.09 -45.82 -59.94 -107.90

EV/FCF (x) -16.83 -52.13 682.39 -304.82 918.39 208.01

P/E (x) -1303.99 -2439.90 -4569.27 -5856.82 -8772.42 -30534.71

P/FCF (x) -12.59 -45.60 671.65 -299.01 901.84 206.82

Free cash flow yield -7.9% -2.2% 0.1% -0.3% 0.1% 0.5%

Dividend yield (%) na na na na na na

EV/NOPAT (x) -14.71 -46.92 -33.73 -50.29 -68.28 -140.83

Source: Company data, Stifel estimates



October 19, 2017

9

OUR CENTRAL CASE

WANdisco is an infrastructure software company that has developed a patent-protected method for data

replication across heterogeneous compute environments. The company focuses mainly on (1) Source

Code Management/aka Application Life Cycle Management and (2) Big Data. WANdisco has been on a

roller coaster since its IPO in 2012, but we believe it has now found its feet with these milestone results.

There is a new coherency with the Fusion product, a better defined sales model, and operating results

look to be turning around.

What does WANdisco do?

WANdisco has built a range of enterprise-class data replication products using its own technology,

DConE. These products improve data migration and ‘round-the-clock’ availability in areas like

replication, mirroring and clustering, and also help to eliminate WAN latency.

At the company’s outset, DConE was focused on a relatively narrow area in version control. The

company expanded into the Hadoop (big) data replication with the 2013 launch of its Non-Stop Name

Node on the heels of its 2012 acquisition of AltoStor. In 2015, WANdisco debuted Fusion, which

significantly expanded TAM as it connected beyond Hadoop distributors into the wider storage market

including vendors such as Amazon, EMC, IBM, Oracle and Teradata. A core technology runs through

the products, which are designed to be independent of the underlying application so they can serve as

the foundation for distributing other applications or databases. WANdisco enables geographically

distributed servers to remain continuously synchronised (i.e. have the same data at the same time). This

solves problems for companies with distributed divisions that are working collaboratively (e.g. in

software design) that operate over a WAN (wide area network). It also appeals to those which are

concerned about network speed, latency, availability, scalability and security. WANdisco competes in

the same markets as infrastructure software companies like CA, IBM, Micro Focus and Microsoft.

How does it do it?

WANdisco has developed an active transaction replication to provide continuous availability, streaming

back-up, uninterrupted migration, hybrid Cloud and Cloud bursting, and data consistency across

clusters that are any distance apart. By this, all the data servers in a network are ‘equal’ (i.e. ‘peers’). To

better understand, compare with active:passive – or better master/slave, where one server is the de

facto controller – should the master go down, data will be lost. So active:active is superior because the

data servers maintain ‘a consensus’ across a distributed network, so the core data remains secure.

WANdisco’s mission

WANdisco aims to push the limits of what can be achieved with distributed computer systems deployed

on a WAN. Note that while WANdisco uses Open Source software and is a member of the Apache

Foundation, DConE itself is not Open Source software. Talking with Dr Yeturu, we gained the

impression that he remains committed to developing software that will create “a richer multipart

interaction with the Internet”. Fundamental to this view, in our opinion, is that the Internet is a distributed

computing platform and so being able to harness various web servers and ‘see’ them as a single entity,

and to coordinate computing resources without using a central server, is crucial.

Who are the customers?

Users are typically larger corporations, and include a number of Fortune Global 100 companies. These

companies are building software, organised globally and migrating to the Cloud. Sectors include auto,

entertainment, financial services, government, healthcare, IT, telecoms and utilities. The customer list

has a strong tech constituency that includes not only Accenture, ARM, Cisco, Dell, HP and IBM, but also

banks like HSBC and such global brands as GE, Fidelity, John Deere, Johnson & Johnson, Pitney

Bowes and Wal-Mart. The company has c200 customers (of which c31 are on WANdisco’s Big Data

product, and most use its ALM product).



October 19, 2017

10

The use case: What do customers use WANdisco for?

Disaster recovery. Customers use WANdisco to ensure that if their server goes down (e.g. the

area electricity gets knocked out), that data would not be lost. WANdisco products provide Cloud,

on-premise and Cloud-to-Cloud with guaranteed data consistency and no data loss.

Data migration to the Cloud. Over time, the Cloud is attracting customers like a moth to a flame,

because it is cost-effective, requires little upfront investment, and brings many other benefits to an

enterprise. This use case will impact many companies, and will be long-running simply because

companies will migrate to the Cloud at different paces. By its nature, however, the Cloud

represents a one-off sale, and arguably less interesting for WANdisco (and its shareholders).

Customers could use a ‘hybrid’ Cloud, where they mix and match on-premise and Cloud servers,

and move data between the two. Other users might move data from the ‘edge’ of the network to

the core. Here a customer talked through a medical example – with on-the-edge data collection

from remote locations and the analysis in the core. This gives WANdisco an annuity revenue

stream.

Inter-Cloud data replication and availability. In this scenario, customers look to maintain dual

supplier strategies (like their on-premise brethren of old), or to migrate to a new Cloud vendor after

losing confidence with the current supplier owing to poor SAL management, predatory pricing and

the cost of extracting data or security issues, etc. They will need to be able to move data between

suppliers in order to avoid predatory pricing and ‘lock-in’. This nascent market currently reflects the

early stage of enterprise users migrating to the Cloud. Some customers have already started to

migrate between Cloud providers, but much of their eagerness to do so gets diluted once they

recognise the downtime required with traditional migration methods. We are reading more about

these Cloud migrations in the trade press suggesting that there is a growing audience.

Improving Cloud provider availability. We note outages at cloud service providers. If data were

being consistently replicated across multiple data centres within a Cloud provider in near-real time,

these outages would not occur. This suggests that data is replicated across cloud data centres in a

batch-based approach – hence the outages.

Impressive ROI

Customers and channel partners speak positively of the realised ROI. While the data are all anecdotal,

a Forrester Total Economic Impact (TEI) study of WANdisco’s Subversion Multisite product revealed a

357% return on investment within a nine-month period.

The sales channel partners

IBM. In April 2016, WANdisco announced an OEM deal with IBM where the latter will resell

WANdisco Fusion under a white-labelled product called IBM Big Replicate. This is a two-year,

nonexclusive deal whereby IBM sells Fusion, offers first-line product support, and WANdisco

provides technical and engineering support to IBM. As to the economics, IBM charges the

customer in an on-premise deployment, and pays a 30% royalty to WANdisco.

Amazon Web Services (AWS). Now signed up as part of AWS ‘Snowball’ initiative (i.e. AWS large

scale data transfer service), this partnership started in March 2016 when WANdisco began selling

on the AWS marketplace, where it was listed as a featured product. We understand that the

economics are an 80/20 split in favour of WANdisco of any revenue generated through AWS. The

pricing for Fusion is a 30% upcharge to S3 pricing. S3 is currently priced at $0.03/GB/Month.

Fusion is 30% on top of this, with WANdisco receiving 80% of that. This is priced on all data that

are under continuous replication. We believe there are four to five live customers via AWS.



October 19, 2017

11

Oracle. WANdisco has a resale agreement in place with Oracle. WANdisco Fusion works with

Oracle’s Big Data Appliance. The first deal through this channel was announced in October 2016,

a US$1.5m deal with an unnamed regional US bank. An Oracle reseller sourced and closed the

deal. As WANdisco is a member of Oracle’s PartnerNetwork, Fusion was certified as an

enterprise-grade solution that met Oracle’s Maximum Availability Architecture best practices

blueprint for replication.

HP. WANdisco announced a resale agreement with HP in Q2 2016. The banner client is the Dubai

Connected City. Given the Micro Focus merger, we would imagine more sales activity in 2017 as

reps get keen to show that they are ‘useful’.

New Context. This partnership was inked in July 2016. New Context is essentially an outsourced

IT management service. The concern that WANdisco addresses here is data integrity for clients

that have data compliance policies, and WANdisco can control where the data goes and who has

access to it in an auditable way. This is important for clients that have classified data with

concomitant rules on read access, and that also have data that are not allowed to leave the US.

Google. WANdisco is a listed technology partner of Google Dataproc. We understand that Google

has developed a solution for data replication for Google search – by using hardware it calls the

Google Spanner database. With this, Google put atomic clocks on all of their servers, launched

GPS satellites, and ran cabling across their data centres. This is deemed to be good enough for

Google search, but is not a commercial solution.

What does Fusion do?

WANdisco’s DConE technology is at the heart of Fusion. It allows multiple instances of the same

application to operate on independent hardware without sharing any resources. This is active

transactional replication technology for continuous availability, streaming backup, uninterrupted

migration and hybrid Cloud, and ensures data consistency across clusters that are any distance apart.

This is possible because all of the application servers are synchronised continuously by the DConE

engine, and operate as peers to each other, regardless of whether the servers are on the same LAN or

are globally separated and accessible only over a WAN. The industry talks about this as being a ‘peer-

to-peer’ system when there is no central co-ordinating ‘master’, or lead service. This is achieved by

immediately replicating changes made against one server to the others (‘active-active replication’).

Using this WANdisco software creates the effect of a single-server system (i.e. a quasi-single instance),

which then performs at LAN speed even though the servers themselves may be thousands of miles

apart.

Once WANdisco’s products are installed at each site, each server becomes an active node on the WAN

with its own DConE. These work cooperatively as peers to perform distributed transaction management

tasks, handle conflicts, and ensure that the same write order is maintained across all of the servers.

This means WANdisco provides One-Copy Equivalence (consistent data/single version of the truth)

across a system of distributed servers connected over a WAN or LAN, and should one server go down

there is no effect on the other servers in the implementation. This has significant implications in terms of

maximising productivity, eliminating downtime and preventing data loss in a globally distributed

collaborative work environment.

If a site goes down, with Fusion installed with each cluster or in each Cloud environment, each cluster

knows the last good transaction it processed. Hence, when it comes back online, it can reach out to the

other servers installed with the other participating clusters, grab all the transactions that it missed while

it was offline, and apply them and re-sync automatically. This eliminates the risk of human error in

recovery, and ensures against data loss. Fusion continuously replicates data.



October 19, 2017

12

RISKS

The key risks to our investment case include: (1) the core technology, (2) the competitive market, (3) the

nature of the demand environment, and (4) WANdisco’s continued ability to deliver further sales growth

at similar-to-recent rates.

Technology risk

In terms of technological risk, we identify two issues:

1. What problem does DConE solve, and is it ‘critical’ enough? Some repositioning has aimed at

‘nailing down’ the use case. However, in WANdisco Fusion, we see a coherent product with a

defined end-market.

2. Is the core technology established? Paxos creator (from his paper ‘The Part-Time Parliament’)

Dr Leslie Lamport acknowledged to us that, despite interest in the Paxos offshoot Raft having led

to a number of implementations, Paxos remains the standard approach to implementing fault-

tolerant systems.

Sidebar: the origins of DConE

The original intellectual property (IP) underpinning DConE was first issued in 2005, after WANdisco’s

technical founder Yeturu Aahlad spent five years working to create a peer-to-peer distributed system. Dr

Aahlad’s work was based on a paper by mathematician/computer scientist Dr Leslie Lamport, who

named his solution the Paxos algorithm (see Appendix). That Dr Aahlad had been a distributed systems

architect at Sun Microsystems tells us that he was in the right place at the right time, when the industry

was just beginning to think about issues around distributed computing.

Good enough and DIY solutions

The technology industry is littered with examples of ‘good enough’ technology (note: this report

was written in Microsoft Word 2010). Good enough often means ‘cheaper’, but also in this case we

have come across examples where companies have developed a (nearly) peer-based system

using batch-processing (i.e. not real-time) and solutions that re-hash the ‘master/slave’

methodology – i.e. they are not peer-to-peer. There is no direct competitor, and the key is data

migration with no interruption (i.e. zero outage).

Some of the ‘2.0’ web properties solve the problem by: (1) using more hardware (the unit of

production in a data centre is a small, and cheap compute blade, easily deployed) by creating a

hardware-based fault tolerance – an easier fix, but not generally suitable; (2) developing their own

solutions. Later in this report, we look at two case studies (Airbnb and Netflix) to illustrate what

cash- and engineering-rich companies can do for themselves.

Raft. Raft is a consensus algorithm designed as an alternative to Paxos. It was meant to be more

understandable than Paxos. Like Paxos, Raft offers a generic route to distributing a state machine

across a cluster of computing systems, ensuring that each node in the cluster agrees upon the

same series of state transitions. The difference is that Raft decomposes the problem into relatively

independent sub-problems. A server in a Raft cluster is either a leader, a candidate, or a follower,

with the leader responsible for log replication and informing the followers via a heartbeat message.

There are a number of implementations (see https://raft.github.io/#implementations), but no large

commercial sponsorship.

Blockchain enjoys some status as a tech cure-all currently, and its distributed ledger technology is

being explored as a trusted way to track the ownership of assets with no need of a central

authority. The design goal is to speed up transactions and cut costs, while lowering fraud

incidences. At its heart, Blockchain is a distributed file system. The database is shared by all

nodes participating in a system, and people using Blockchain keep copies (a block) of the



October 19, 2017

13

Wandisco Blockchain file. As such, the Blockchain database uses a distributed consensus model.

Each block has a cryptographic signature (aka a hash) of the preceding block, hence the ‘chain’

analogy as the blocks are added sequentially. It is a peer-to-peer network with many distributed

nodes. While the failure of one node (or even several nodes) will not prevent the rest of the

network from functioning properly, data will still be lost. Current criticisms of Blockchain include its

(lack of) speed and the visibility of the information (to anyone). We know of one WANdisco

customer that had been exploring Blockchain, but now sees DConE as more suitable for its needs.

Artificial intelligence taking over

Just to future-proof this report – there are moves afoot to bring more artificial intelligence (AI) into the

software development world. It is not too much of a stretch, should we accept it, that AI develops

software autonomously. In that case, ‘out of the box’ fault tolerance could be at such a level that,

coupled with AI developed software, knowledge (i.e. the practical implementation of data) will be

ubiquitous and will always be available. This would negate the need for DConE. We offer no timeframe.

Running out of cash

WANdisco has returned to seek cash from the stock market several times. The latest trading update (16

January 2017) revealed that in Q4 the cash burn was US$200k, down from US$6.9m in 2016. Among

several corners the company has turned: (1) it has recut its cloth to match existing cash resources, and

(2) a new appreciation of the importance of the balance sheet, a sea change we think was inspired by

the new CFO. Our model suggests that the company does not need to raise fresh cash, but this

depends on rising revenue and better sales execution, rather than trimming the cost base.

Sales execution is there at last

As befits an early-stage software company, WANdisco has had a multifaceted ‘let’s try a few things’ go-

to-market strategy. We recall its pre-IPO days, when WANdisco built a ‘frictionless’ sales model for its

ALM line of business, supplemented with an ‘inside’ sales team that concentrated on converting the free

community to paying customers. However, ALM/version control customers were a technical audience

that knew what it wanted. The same approach was never going to work for Big Data, where WANdisco

debuted in 2013, and where it would have to build its own enterprise sales team supplemented by a

sales channel. The early moves with the Hadoop ‘distro’ companies (Cloudera and Hortonworks) looked

sensible, yet ultimately proved to be the wrong start point as the distros developed their own Disaster

Recovery (DR) offer. Through iterations, WANdisco now has a multifaceted sales distribution model that

headlines with a rare IBM OEM relationship. In addition, WANdisco has its own enterprise sales team,

coupled to channel partners including Amazon and Oracle, plus a number of professional services

organisations.

Competition – a wrinkle

Looking through the competitive pack (see below), we note a product competitor in the ALM market,

‘Git’. Git is an Open Source, distributed version control system designed to handle all projects, from

small to very large, with speed. It is positioned as a replacement product for version control tools like

Perforce or Subversion. Git has surged in recent years, and companies like GitHub and (to some extent)

Atlassian have done a better job a tethering themselves to the Git banner. While WANdisco has a Git

(and indeed a Gerrit) offer in addition to Subversion, arguably the company needs to work harder to

establish its brand in the Git market.

‘Forking’

WANdisco is an engineering-led software company that puts technology on a pedestal, and its

customers look to it to figure out ‘what’s next’. In such situations, ‘forking’ comes with the territory in the

Open Source world. ‘Forking’, or the development of competing variations, occurs when projects

splintering into different forms lead to disputes. However, a number of recent staff and process changes

in engineering at WANdisco should help reduce the risk of forking.



October 19, 2017

14

Difficulties in hiring

As a small company with developers in San Francisco, WANdisco may find it difficult to hire. In fairness,

given that options are ‘under the water’, we are surprised that unplanned staff attrition is not more of an

issue. We caution that:

Good people remain ‘hard’ to hire;

The US, Northern Ireland and Sheffield offices may find it difficult to hire. In the US, the staff may

be too ‘footloose’ and expensive, too prone to receiving competitor calls. In Sheffield, they may be

too scarce. In Northern Ireland, they might not be skilled enough. We like the global nature of

WANdisco, and think the company needs to have a distributed office structure to mirror how its

customers are organised.

Staff are expensive

We refer to ITJobswatch for an impression of UK software developer costs for contractor and full-time

staff. Staff costs are c70% of any tech company’s operating spend. Most of WANdisco’s staff are in

three locations. Of these, we note that a Hadoop developer in the UK is £800-850/day for a contractor,

c£64,000 per annum full time. This is a UK average, and local rates in Sheffield will be cheaper than the

same role in the City of London. The same role in San Francisco will average US$112,000 for FTE –

that translates to £90,000, or c40% more on a ceteris paribus basis. While the issue of staff costs will be

food for thought, balancing this we recall a conversation with analysts at another tech company, who

noted that there was a big difference in two hypothetical developers just on the basis of the influence of

the surroundings, and suggested that developers are better in somewhere like Silicon Valley, as it is a

magnet for new thought and best practice.



October 19, 2017

15

Competitive overview

Signing agreements with key players in the space (IBM, Oracle, Amazon and Microsoft) all indicate that

(i) WANdisco IP is extremely broad, (ii) there is a general acknowledgement that there isn’t anyone

doing anything similar, and in their view (iii) WANdisco has a multi-year head-start over anyone. There

is no direct competitor. However, looking into the market we see a number of competitive areas:

We can see operational overlaps with ETL (Extract, Transform, Load) vendors in the data

migration business. The long list of established companies focused on this area includes

Informatica, Information Builders, SAS and Syncsort. Here, the competitive threat can be

somewhat overplayed.

There is also some overlap with the database replication vendors. This is more often a

master/slave architecture between the original and the copies. The master logs the updates, which

then ripple through to the slaves which output a message, stating that it has received the update

successfully, thus allowing the sending (and potentially re-sending until successfully applied) of

subsequent updates. In the case of multi-master replication, the updates can be submitted to any

database node, and then ripple through to other servers – but here the challenge is transactional

conflict prevention. Database replication also becomes difficult when it scales up.

In addition, there are more niche Big Data replication vendors like Attunity, Denodo, Talend and

many others. Looking at (e.g. Attunity), we note that its Replicate software helps load and ingest

data across all major databases, data warehouses and Hadoop, on-premises and in the Cloud.

Attunity’s customer base for this product set is estimated to be c2,000 organisations globally.

Through capabilities for replicating data to and from the Cloud, and accelerating data warehouse

deployments, Big Data (Hadoop) and workload optimisation support, Attunity enables a broad

scope of integration styles, and also supports Apache Kafka by enabling real-time intertwining of

data movement.

The most potent are the IaaS and PaaS vendors. Here we find the solution is first to put hardware

server technology at the problem. Note for example that Amazon Web Services has disaster

recovery. But looking at the offer, we note Amazon has an AWS Import/Export feature when users

are moving large amounts of data into and out of AWS – it is portable storage device, which for

data sets of significant size, AWS Import/Export is often faster than Internet transfer. For data

migration, as customers move to the Cloud, AWS offers Amazon Relational Database Service,

alongside Amazon DynamoDBis (a NoSQL database) and Amazon Redshift is a fully managed,

petabyte-scale data warehouse service.

Other Cloud infrastructure software vendors, such as distro Cloudera or Hortonworks. We have

seen distros and infrastructure Hadoop companies launch their own, mostly, Disaster Recovery

based solutions. For example, Cloudera Manager Backup and Disaster Recovery (BDR) is its

Hadoop DR offer. These are typically active:passive and likely reflect early Hadoop usage and the

use of non-mission critical applications. On user forums, we find lots of interest in using the

‘snapshot feature’ in HDFS. These will take a point in time image, which in truth could be for an

entire file system, a sub-tree in a file system or just a file. Clearly this will not capture the

incremental data changes, and snapshotting can be bandwidth-intensive.

Other technology solutions – here, we have Raft and to some extent Blockchain, yet neither

has a major commercial endorsement. Docker is also lurking in the background. While not directly

competitive, it allows users to develop distributed computing systems by using software

‘containers’. Docker can be integrated into various infrastructure tools, including Amazon Web

Services, Ansible, CFEngine, Chef, Google Cloud Platform, IBM Bluemix, HPE Helion, Microsoft

Azure, OpenStack, Oracle Container Cloud Service and others. According to industry analysts 451

Research, ‘Docker is a tool that can package an application and its dependencies in a virtual



October 19, 2017

16

container that can run on any Linux server. This helps enable flexibility and portability on where the

application can run, whether on premises, public Cloud, private Cloud, bare metal, etc.’ Kafka also

deserves a mention. Kafka is used for building real-time data pipelines and streaming applications.

This is an Apache Open Source project. Kafka lets users publish and subscribe to streams of

records so it is a messaging system. Kafka also lets users store streams of records in a fault

tolerant way. So we can see advantages like scalability and fault-tolerance so there is some

overlap with WANdisco. Kafka is also widely used and it may become more popular as users look

to ingest data into a Hadoop cluster and then cleanse it. (Consider a use case: you might trawl the

web for all references for all daily commuters into the City of London – use Kafka to gather that raw

data. Then run an analytics to find the only Irish guy with five kids.)

DIY. (1) Airbnb posted a blog detailing its migration of a Hive warehouse. The warehouse had

grown from 350TB in mid-2013 to 11PB by the end of 2015. The sheer size of the warehouse gave

rise to issues of reliability, so Airbnb opted to migrate to a new architecture. Airbnb decided that

existing migration tools either had issues with a large data warehouse or had a significant

operational overhead, so it developed ‘ReAirto’ to save time and effort when replicating data at this

scale. Initially, all the data was in a single HDFS/Hive warehouse, but the mixed production with ad

hoc workloads led to reliability issues. Airbnb split the warehouse in two: production jobs and ad

hoc queries. It then had to migrate the large warehouse, and then after the split keep datasets in

sync. It developed ‘ReAir’ to do this, and has since open-sourced the tool for the community.

‘ReAir’ is useful for replicating data warehouses based on the Hive metastore, and Airbnb has

made these tools to be scalable to clusters which are petabytes in size. ReAir can work across a

variety of different Hive and Hadoop versions and can operate largely standalone.

ReAir includes two replication tools: batch and incremental.

A batch replication tool to copy a specific list of tables at once, which is ideal for a cluster

migration.

An incremental replication tool to track changes that occur in the warehouse and replicate the

objects as they are generated or modified. This keeps datasets in sync between clusters as it

starts copying changes within seconds of object creation.

Link to GitHub: https://github.com/airbnb/reair

(2) Netflix has a loosely coupled microservice-based architecture that emphasises separation of

concerns. It has developed EVCache, its data-caching service that provides the low-latency, high

reliability caching solution, and it routinely handles upwards of 30m requests/sec, storing hundreds

of billions of objects across tens of thousands of memcached instances, translating to c2 trillion

requests a day globally. When Netflix launched in 130 additional countries, it built EVCache’s

global replication system. EVCache is Open Source, and has been in production for more than five

years. Netflix’s Cloud-based service is spread across three Amazon Web Services (AWS) regions.

While requests are mostly served from the region the member is closest to, this can change due to

(say) problems with infrastructure or a regional/geographic failover, so because of this Netflix has a

stateless application server architecture, whereby a server can take a request from any region.

Consequently the data must be replicated to all regions so it’s available to serve member requests

no matter where they originate. Netflix designed EVCache for itself, so it admits that one non

requirement was to have strong global consistency. Remember that Netflix doesn’t mind if (say)

the Ireland and Virginia servers have slightly different recommendations for the same person as

long as the difference does not impact browsing or streaming experience. For non-critical data,

Netflix has an “eventual consistency” model for replication, whereby local or global differences are

tolerated for a short time. This simplifies the EVCache replication design, and means that Netflix

did not worry about global locking, quorum reads and writes, transactional updates, partial-commit

rollbacks or other complications of distributed consistency.



October 19, 2017

17

Figure 5: Competitive overview

Source: Stifel Research

Figure 6: Profit & loss (US$m)

Year to 31 December 2015A 2016A H1 2017E 2018E 2019E 2020E

Bookings 9.00 15.50 10.20 20.83 29.29 40.60 50.35

SCM/ALM 6.50 8.40 3.20 8.28 7.88 7.64 7.54

Fusion/Big Data 2.50 7.10 7.00 12.56 21.41 32.96 42.81

Turnover by type

SCM/ ALM 9.2 8.2 4.6 8.4 8.0 7.8 7.7

Fusion/big data 1.8 3.2 5.1 8.4 12.1 16.5 21.4

Total turnover 11.0 11.4 9.66 16.8 20.1 24.3 29.1

Adj EBITDA pre exceptional -15.988 -7.46 0.25 -2.41 0.02 2.84 6.74

Margin (%) -145% -0.66 0.03 -14% 0% 12% 23%

Exceptional costs -0.614 8.11 -2.30 -2.30 0.00 0.00 0.00

Adj EBITDA post exceptional -16.602 0.65 -2.04 -4.71 0.02 2.84 6.74

EBITDA margin % -151% 6% -28% 0% 12% 23%

Adj EBITDA pre exceptional inc Dev expenditure -24.357 -13.324 -2.77 -7.90 -5.86 -3.56 0.11

Depreciation -0.270 -0.17 -0.11 -0.20 -0.22 -0.24 -0.27

Adj EBITA -16.258 -7.638 0.15 -2.62 -0.20 2.60 6.47

Net Interest -0.51 -0.27 -0.24 0.00 0.00 0.00 0.00

Adjusted pre tax profit -16.764 -7.91 -0.09 -2.62 -0.20 2.60 6.47

Exceptional items -0.614 8.11 -2.30 -2.30 0.00 0.00 0.00

Share based payments -4.057 -1.787 -0.46 -1.25 -1.38 -1.51 -1.66

Intangible amortisation -9.600 -8.466 -3.44 -7.62 -8.00 -8.40 -8.82

Pre-tax profit post exceptional -31.035 -10.046 -6.288 -13.78 -9.58 -7.32 -4.01

Taxation 1.13 0.772 -0.04 0.81 0.85 0.89 0.94

Profit after tax -29.906 -9.274 -6.33 -12.97 -8.73 -6.42 -3.08

Other income 0.055 0.107 0.00 0.11 0.12 0.12 0.13

Dividends 0.0 0.0 0.00 0.00 0.00 0.00 0.00

Reported retained profit post exceptional -29.851 -9.167 -6.33 -12.86 -8.61 -6.30 -2.95

Weighted average basic shares (m) 26.78 33.29 37.32 37.66 37.66 37.66 37.66

Weighted average full diluted shares (m) 28.78 33.29 37.32 37.66 37.66 37.66 37.66

Diluted adjusted EPS (US$) -0.88 -0.47 -0.10 -0.25 -0.20 -0.13 -0.04

DPS (US$) 0.0 0.0 0.00 0.0 0.0 0.0 0.0




October 19, 2017

18

Figure 7: Cash flow (US$m)


Operating Profit -29.91 -9.27 -12.97 -8.73 -6.42 -3.08

Depreciation & Amortisation 9.87 8.64 7.82 8.22 8.64 9.09

Share based payments 4.67 1.82 1.25 1.38 1.51 1.66

Increase/(decrease) creditors/payables -0.43 0.83 0.75 0.44 0.31 0.59

Increase/(decrease) deferred income -1.51 2.74 2.36 3.15 2.72 0.00

(Increase)/decrease debtors/receivables 0.28 0.33 -0.53 -0.75 -0.88 -0.45

(Increase)/decrease in Gov grant -0.05 -0.01 0.00 0.00 0.00 0.00

Working capital -1.71 3.88 2.57 2.84 2.14 0.14

Other 0.04 -7.24 6.00 0.00 0.00 0.00

Operating cash flow -17.04 -2.18 4.67 3.71 5.88 7.81

Net interest -0.06 -0.17 0.16 0.15 0.13 0.12

Taxation -0.58 -0.08 0.81 0.60 0.63 0.66

Net capex -0.10 -0.06 -0.07 -0.07 -0.07 -0.08

Development -8.37 -5.86 -4.94 -5.82 -6.08 -6.43

Free cash flow -26.14 -8.35 0.64 -1.44 0.48 2.08

Dividends 0.00 0.00 0.00 0.00 0.00 0.00

Acquisitions 0.00 0.00 0.00 0.00 0.00 0.00

Other 0.06 0.00 0.00 0.00 0.00 0.00

Net cash flow -26.08 -8.35 0.64 -1.44 0.48 2.08

Shares issued (net) 26.17 13.52 0.00 0.00 0.00 0.00

Cash/(debt) acquired 0.00 0.00 0.00 0.00 0.00 0.00

Currency effects -0.02 -0.18 0.00 0.00 0.00 0.00

Increase / (decrease) cash 0.07 5.00 0.64 -1.44 0.48 2.08

Opening cash / (debt) 2.48 2.56 7.56 8.20 6.76 7.23

Closing cash / (debt) 2.56 7.556 8.20 6.76 7.23 9.32

Borrowings/finance lease 0.00 0.29 0.31 0.34 0.36 0.00

Other loans 0.00 0.00 3.00 3.00 3.00 0.00

Net cash (debt) 2.56 7.26 4.89 3.42 3.88 9.32




October 19, 2017

19

Figure 8: Balance sheet (US$m)


Fixed Assets

P. P & Equipment 0.23 0.49 0.44 0.38 0.41 0.37

Intangible assets 8.58 5.98 4.84 3.89 4.47 4.03

Total fixed assets 8.81 6.47 5.29 4.27 4.89 4.41

Current Assets (Receivables)

Cash at hand and in bank 2.56 7.56 8.20 6.76 7.24 9.32

Trade debtors 6.73 6.15 6.68 7.43 8.32 8.77

Other receivables & prepayments 0.00 0.00 0.00 0.00 0.00 0.00

Corp tax credit receivable 0.00 0.00 0.00 0.00 0.00 0.00

Total current assets 9.28 13.70 14.88 14.19 15.55 18.09

Total assets 18.10 20.17 20.17 18.46 20.44 22.50

Current liabilities (Payables)

Short-term debt 0.00 0.09 0.09 0.09 0.09 0.09

Trade creditors 2.71 3.49 4.24 4.68 4.99 5.58

Deferred income 6.06 5.81 8.17 11.32 14.04 14.04

Current tax liabilities 0.00 0.01 0.00 0.00 0.00 0.00

Deferred government grant 0.03 0.01 0.01 0.01 0.02 0.02

Total current liabilities 8.80 9.41 12.51 16.11 19.13 19.73

Net Current Assets 0.48 4.29 2.37 -1.91 -3.58 -1.64

Total Assets-Current Liabilities 9.29 10.76 7.66 2.36 1.31 2.77

Non-current liabilities

Deferred income 3.70 6.68 8.98 12.63 17.51 21.71

Borrowings/finance lease 0.00 0.29 0.31 0.32 0.34 0.36

Borrowings 3rd pty loan 3.00 3.00 3.00 3.00

Deferred tax liability 0.01 0.00 0.10 0.11 0.12 0.13

Retirement benefit obligations 0.00 0.00 0.00 0.00 0.00 0.00

Total non-current liabilities 3.70 6.98 12.39 16.07 20.97 25.20

Total Liabilities 12.50 16.39 24.90 32.17 40.11 44.93

Net assets 5.59 3.78 -4.74 -13.71 -19.67 -22.44

Shareholders’ Funds

Called up share capital 4.67 5.64 5.92 6.22 6.53 6.85

Share premium 81.97 94.53 94.53 94.53 94.53 94.53

Translation reserve -0.25 -8.28 0.15 2.56 3.58 4.89

Other reserves 1.25 1.25 1.25 1.25 1.25 1.25

Retained earnings -82.05 -89.34 -106.58 -118.26 -125.54 -129.96

Shareholders’ funds 5.59 3.78 -4.74 -13.71 -19.66 -22.44

Equity and liabilities 18.10 20.17 20.16 18.46 20.44 22.49




October 19, 2017

20

Important Disclosures and Certifications

I, George O'Connor, certify that the views expressed in this research report accurately reflect my personal viewsabout the subject securities or issuers; and I, George O'Connor, certify that no part of my compensation was,is, or will be directly or indirectly related to the specific recommendations or views contained in this researchreport. Our European Policy for Managing Research Conflicts of Interest is available at www.stifel.com.

WANdisco plc (WAND.LN) as of October 18, 2017 (in GBp)

Pric

e (G

Bp)

1,000

800

600

400

200

0Oct-14 Jan-15 Apr-15 Jul-15 Oct-15 Jan-16 Apr-16 Jul-16 Oct-16 Jan-17 Apr-17 Jul-17 Oct-17 Jan-18

02/23/2017390.50I:B:508.89p

03/08/2017480.00B:606.66p

04/24/2017462.00B:622.70p

09/06/2017700.00B:842.74p

09/26/2017817.50B:843.00p

Powered by: BlueMatrixBuy=B; Hold=H; Sell=S; Discontinued=D; Suspended=SU; Discontinued=D; Initiation=I*Represents the value(s) that changed.

For a price chart with our ratings and any applicable target price changes for WAND.LN go tohttp://stifel2.bluematrix.com/sellside/Disclosures.action?ticker=WAND.LN

Stifel or an affiliate is a market maker or liquidity provider in the securities of WANdisco plc.Stifel or an affiliate has received compensation for investment banking services from WANdisco plc in the past 12 months.Stifel or an affiliate expects to receive or intends to seek compensation for investment banking services from WANdiscoplc in the next 3 months.WANdisco plc is provided with non-investment banking, securities related services by Stifel or an affiliate or was providedwith non-investment banking, securities related services by Stifel or an affiliate within the past 12 months.WANdisco plc is provided with investment banking services by Stifel or was provided with investment banking servicesby Stifel or an affiliate within the past 12 months.WANdisco plc is a client of Stifel or an affiliate or was a client of Stifel or an affiliate within the past 12 months.Stifel or an affiliate has received compensation for non-investment banking, securities related services from WANdiscoplc in the past 12 months.Stifel or an affiliate is a corporate broker and/or advisor to WANdisco plc.The equity research analyst(s) responsible for the preparation of this report receive(s) compensation based on variousfactors, including Stifel’s overall revenue, which includes investment banking revenue.Our investment rating system is three tiered, defined as follows:

BUY -We expect a total return of greater than 10% over the next 12 months with total return equal to the percentageprice change plus dividend yield.

HOLD -We expect a total return between -5% and 10% over the next 12 months with total return equal to the percentageprice change plus dividend yield.

SELL -We expect a total return below -5% over the next 12 months with total return equal to the percentage price changeplus dividend yield.

Occasionally, we use the ancillary rating of SUSPENDED (SU) to indicate a long-term suspension in rating and/or targetprice, and/or coverage due to applicable regulations or Stifel policies. SUSPENDED indicates the analyst is unable todetermine a “reasonable basis” for rating/target price or estimates due to lack of publicly available information or the



October 19, 2017

21

http://stifel2.bluematrix.com/sellside/Disclosures.action?ticker=WAND.LN

inability to quantify the publicly available information provided by the company and it is unknown when the outlook willbe clarified. SUSPENDED may also be used when an analyst has left the firm.

Of the securities we rate, 49% are rated Buy, 38% are rated Hold, 2% are rated Sell and 11% are rated Suspended.

Within the last 12 months, Stifel or an affiliate has provided investment banking services for 20%, 8%, 0% and 10% ofthe companies whose shares are rated Buy, Hold, Sell and Suspended, respectively.

Additional Disclosures

Please visit the Research Page at www.stifel.com for the current research disclosures and respective target pricemethodology applicable to the companies mentioned in this publication that are within Stifel's coverage universe. For adiscussion of risks to target price please see our stand-alone company reports and notes for all stocks.

The information contained herein has been prepared from sources believed to be reliable but is not guaranteed by usand is not a complete summary or statement of all available data, nor is it considered an offer to buy or sell any securitiesreferred to herein. Opinions expressed are subject to change without notice and do not take into account the particularinvestment objectives, financial situation or needs of individual investors. Employees of Stifel, or its affiliates may, attimes, release written or oral commentary, technical analysis or trading strategies that differ from the opinions expressedwithin. Past performance should not and cannot be viewed as an indicator of future performance.

As a multi-disciplined financial services firm, Stifel regularly seeks investment banking assignments and compensationfrom issuers for services including, but not limited to, acting as an underwriter in an offering or financial advisor in a mergeror acquisition, or serving as a placement agent in private transactions.

Affiliate Disclosures

“Stifel”, includes Stifel Nicolaus & Company (“SNC”), a US broker-dealer registered with the United States Securitiesand Exchange Commission and the Financial Industry National Regulatory Authority and Stifel Nicolaus Europe Limited(“SNEL”), which is authorized and regulated by the Financial Conduct Authority (“FCA”), (FRN 190412) and is a memberof the London Stock Exchange.

Registration of non-US Analysts: Any non-US research analyst employed by SNEL contributing to this report is notregistered/qualified as a research analyst with FINRA and is not an associated person of the US broker-dealer andtherefore may not be subject to FINRA Rule 2241 or NYSE Rule 472 restrictions on communications with a subjectcompany, public appearances, and trading securities held by a research analyst account.

Country Specific and Jurisdictional Disclosures

United States: Research produced and distributed by SNEL is distributed by SNEL to “Major US Institutional Investors”as defined in Rule 15a-6 under the US Securities Exchange Act of 1934, as amended. SNC may also distribute researchprepared by SNEL directly to US clients, including US clients that are not Major US Institutional Investors. In theseinstances, SNC accepts responsibility for the content. SNEL is a non-US broker-dealer and accordingly, any transactionby a US client in the securities discussed in the document must be effected by SNC. US clients wishing to place an ordershould contact their SNC representative.

Canadian Distribution: Research produced by SNEL is distributed in Canada by SNC in reliance on the internationaldealer exemption. This material is intended for use only by professional or institutional investors. None of the investmentsor investment services mentioned or described herein is available to other persons or to anyone in Canada who is not a“permitted client” as defined under applicable Canadian securities law.

UK and European Economic Area (EEA): This report is distributed in the EEA by SNEL, which is authorized andregulated in the United Kingdom by the FCA. In these instances, SNEL accepts responsibility for the content. Researchproduced by SNEL is not intended for use by and should not be made available to non-professional clients.

The complete preceding 12-month recommendations history related to recommendation(s) in this research report isavailable at https://stifel2.bluematrix.com/sellside/MAR.action

Brunei: This document has not been delivered to, registered with or approved by the Brunei Darussalam Registrar ofCompanies, Registrar of International Business Companies, the Brunei Darussalam Ministry of Finance or the AutoritiMonetari Brunei Darussalam. This document and the information contained within will not be registered with any relevantBrunei Authorities under the relevant securities laws of Brunei Darussalam. The interests in the document have not beenand will not be offered, transferred, delivered or sold in or from any part of Brunei Darussalam. This document and the



October 19, 2017

22

information contained within is strictly private and confidential and is being distributed to a limited number of accreditedinvestors, expert investors and institutional investors under the Securities Markets Order, 2013 ("Relevant Persons")upon their request and confirmation that they fully understand that neither the document nor the information containedwithin have been approved or licensed by or registered with the Brunei Darussalam Registrar of Companies, Registrar ofInternational Business Companies, the Brunei Darussalam Ministry of Finance, the Autoriti Monetari Brunei Darussalamor any other relevant governmental agencies within Brunei Darussalam. This document and the information containedwithin must not be acted on or relied on by persons who are not Relevant Persons. Any investment or investment activityto which the document or information contained within is only available to, and will be engaged in only with RelevantPersons.

In jurisdictions where Stifel is not already licensed or registered to trade securities, transactions will only be affected inaccordance with local securities legislation which will vary from jurisdiction to jurisdiction and may require that a transactionis carried out in accordance with applicable exemptions from registration and licensing requirements. Non-US customerswishing to effect transactions should contact a representative of the Stifel entity in their regional jurisdiction except wheregoverning law permits otherwise. US customers wishing to effect transactions should contact their US salesperson.

The recommendation contained in this report was produced at 19 October 2017 01:47EDT and disseminated at 19October 2017 01:47EDT.

The securities discussed in this report may not be available for sale in all jurisdictions and may have adverse taximplications for investors. Clients are advised to speak with their legal or tax advisor prior to making an investmentdecision.

Additional Information is Available Upon Request

© 2017 Stifel. This report is produced for the use of Stifel customers and may not be reproduced, re-distributed or passedto any other person or published in whole or in part for any purpose without the prior consent of Stifel. Stifel NicolausEurope Ltd. 150 Cheapside, London, EC2V 6ET. Registered in England Number 03719559



October 19, 2017

23

trend is that the wandisco plc october 19, 2017 · wandisco cmd; veni, vidi, cc decies centena cmd...

Documents