Enterprise Vault and KVS 1997-2004
Curator’s Note: Nigel Dutt, a successful software engineer-turned-entrepreneur, tells the story of the first commercial email archiving product which today still is one of the leading products–RE
Before KVS: 1997-1999
Conception
The genesis of Enterprise Vault was early in 1997. At that time Digital had decided to cut back investment in its own office product, ALL-IN-1 OfficeServer running on Open VMS, in favour of putting its weight behind Microsoft’s Exchange. Although of course this was not a happy proposition for the ALL-IN-1 engineering group based in Reading, it was arguably sensible for Digital in the long term. The plan was for the ALL-IN-1 team to be dramatically reduced following a major release and for the engineers to be moved to a yet to be defined new product that was to be some sort of add-on for Exchange installations. The original product we proposed was a large scale, searchable store for unstructured information, harnessing Digital’s own Altavista search and various Digital storage technologies. As originally presented it had various target scenarios including email archiving; however, one of our product marketeers, who had just returned to Digital from a role as an early Exchange manager opined that the email archiving target would be worth pursuing in its own right. So in May we made the “big pitch†to the relevant funding SVP with a product called “The Information Warehouseâ€.
Fig. 1: Title page of 'The Information Warehouse" presentation
The proposition was for a central, searchable, scalable and cost-effective archive repository for email collected online from multiple email servers, the initial target being Exchange. The target drivers were storage management, regulatory retention and knowledge preservation. We got the go-ahead for this and the development project started, using the engineers now freed up from ALL-IN-1 3.2, who all retrained for the NT (later to become Windows) platform. Quite early in the project we renamed the product to “Digital Enterprise Vaultâ€. A myth has occasionally arisen that somehow Enterprise Vault was derived from, or was even a rebadging of ALL-IN-1, but it was a ground-up brand new development on the Windows platform (then Windows NT) rather than OpenVMS. The confusion probably arose because the engineering team had all come from ALL-IN-1. However, what we did carry forward as an engineering team was the sort of process-driven engineering methodology that we had developed for ALL-IN-1 engineering (following the SEI Capability Maturity Model) as well as building translatability into the product from the start.
Birth
After the usual torrid software development cycle and quite a long struggle to get some intransigent bugs out of the major beta test at Goldman Sachs, the product was finally released to manufacturing late on Christmas Eve 1998, if for no other reason than that the engineering team wanted to join their families for the holidays.
Fig.2: Original Graphic for Digital branded Enterprise Vault V1.0
By the time of shipping, Compaq had acquired Digital and so the product actually went out as “Compaq Enterprise Vault V1.0â€. Initially the Vault only addressed mailbox archiving because Exchange journaling hadn’t shipped at the time, but that major feature was added to the Vault in a point release shortly afterwards, once the Exchange feature had shipped. During this time we also designed a spin-off product called “Enterprise Vault Mailbox Manager†(VMM). This was done in cahoots with Microsoft and was to be a “mailbox janitor†that cleaned out mailboxes subject to flexible policies. Microsoft had provided a “quick ‘n dirty†add-on to do this for the then current version of Exchange and it had proved extremely popular, so they wanted a “properly engineered†version of it for the next version but did not have the in-house resources to do it themselves alongside their work on the main product. Effectively it was a cut-down client and policy administration tool for Enterprise Vault but without the back-end archiving services, so older messages could be deleted but not saved. The agreement with Microsoft was that VMM would be an optional add-on to Exchange V6.0 and would accompany that product when it shipped later in 2000. We would supply it for free but, in Microsoft’s words, we would have “full bragging rights†to then promote the full Compaq archiving product to the customers. Of course this was the marketers’ dream – access to 100% of our target market with the distribution all done via Microsoft. What could go wrong?
Fig.3: Box cover for Compaq Enterprise Vault V1.0

Fig.4: Screenshot of un-released "Volume Manager" product
Eclipse
By August 1999, two point releases of the Vault had been shipped, twenty or so sales had been made and a number of pilots were under way. Then, following the usual early warning signs as Compaq cut down on the software side of the business, the product was cancelled and the whole engineering team was made redundant. The very next day we all stood outside and watched a total eclipse of the sun, which seemed somehow symbolic at the time. While most of the group went off on their gardening leave, a few of us stayed behind to produce an “unarchive†tool to enable the existing customers to restore everything back to Exchange from their Vaults.
Re-birth
While this was going on, and since the investment climate at the time was very good, we decided to pitch the Vault as a viable basis for a new software company. At the time the Vault was the only purpose-built email archiving product shipping, the nearest thing to a competitor being the UNIX-based Storagetek product, so we felt we had a strong proposition. One of the first people approached was Durlacher, who were at the time the hottest thing in technical venture funding in the UK. They immediately went for the proposal and came up with the initial tranche of money. So, almost before we knew what had happened, we were starting a company. On November 9th 1999 kVault Software (KVS) was formed; the “k†being for “knowledge†which at the time was just supplanting “e†as the trendy company name prefix of choice.
Fig.5: KVS logo
The company started with twenty two of the redundant engineers and two administrators. Nigel Dutt (CTO), Eileen Christie (VP Engineering) and Derek Allan (product architect) continued with the roles they had had in Compaq (and a lot more besides) and Edward Forwood of Durlacher was acting CEO. Unfortunately, Compaq played hardball and set a take-it-or-leave-it price of several million dollars for the IP, to be paid over four years, unlike other employee spin-offs who had only been charged a nominal amount. However, they were well motivated to please the existing customers, so when we started KVS we were hosted in the same desks at Compaq that we had just left, albeit with special KVS badged entry cards. On February 14th 2000 we shipped KVS Enterprise Vault V2.0 from those premises and shortly afterwards we moved to our own address at Winnersh Triangle, near Reading, and finally felt like a “proper companyâ€. As an aside, when looking for a home for us and the product in 1999 I approached Veritas, who replied “no thanks, we are planning to do our own thing†– something I reminded them of when they acquired us in 2004!
KVS: 2000-2004
Early Days at KVS
Fig.6: KVS Enterprise Vault manual cover 2000
Unsurprisingly the customer pipeline that had been established at Compaq mostly evaporated as KVS started up. Instead of buying from a major multi-national, the prospective customers were being asked to put their faith in a small British-based start-up, and for the most part that wasn’t about to happen. However, we did make some very early sales, the first being to Libertel in the Netherlands who purchased in our first week of operation. They were followed as early adopters by Freshfields in the UK, Clearstream in Luxemburg, Shaw Cable in Canada and Jackson National Life in the USA, which meant that we were instantly a multi-national! These early sales were all via Compaq, except for Freshfields that was a Resolution Systems deal. In our first month we also partnered with Essential Computing who we knew well from our Digital days and they organised our KVS launch in London in December 1999. This event then led to more early business through Essential. Similarly, because of the uncertainty when we were first cancelled by Compaq, the VMM project with Microsoft fell through, so we missed what would have been a fabulous opportunity for a small company to kick-start the business. We set up a technical support and sales presence in the USA almost immediately and our US HQ was eventually established in Arlington (Texas) if for no other reason than that was where our North Americas manager lived. Over time several home-based regional offices were set up around North America, in particular in the New York area, reflecting the strong financial sector business we were doing there. Having started with the ideal situation (ideal for engineers that is) of being an engineer-only company, we and our investors knew that we would need to hire a CEO and then develop the sales and marketing functions, and that started to happen after a few months when we hired Mike Hedger as CEO and he in turn recruited the people to start the various sales and other field organisations.
KVS growth
Over its almost five years of life, KVS grew rapidly and consistently in spite of the problems of the dot com crash of 2000-2002.
- Revenue: in our first full year, 2000, revenue was just over $1m, and it grew consistently to $23m in 2003 and just before acquisition in September 2004 the quarterly revenue was $10m, implying an annual run rate of $40m. Over this time, as expected, support, training and consultancy revenue all grew on top of license revenue and were contributing some 30% by 2004. The direct sales/indirect-sales-via-channel revenue split was always fairly consistent at 50/50.
- People: over its lifetime the company grew from just over 20 UK-based engineers to around 220 people covering multiple functions worldwide. Reading in England remained the HQ, where around half the company was based, including the entire engineering group. We had about 75 people in the USA and 35 in the various EMEA and Australian sales offices.
- Geographic Reach: as already mentioned, we sold in multiple countries from the start. We initially sold only English language versions of the product, but translated versions soon followed, with European first and later Asian languages. We aimed to have a US presence right away and the North American sales organisation was built up just behind the EMEA organisation and was around the same size. French and German offices were started early followed later by Australia, Netherlands and Denmark. The Australian office dealt with Asia/Pacific in general. Roughly speaking there was a 40%/55% revenue split between USA/Europe with around 5% from the rest of the world. Reflecting that it was our base country, UK always contributed disproportionately well.
- Funding: we started the company with around twenty five well paid people in nice new premises, so our cost run-rate was significant from Day One. It helped that our office furniture was loaned to us for free by Compaq (in fact it was our familiar old Digital office furniture that had recently been replaced by Compaq and which, like ourselves, was redundant) and all of our Compaq computers were sold to us as a job lot. We quickly learned about the delay between orders being closed and invoices being paid, so we went through a cash flow problem early on and a couple of us founders even paid the salary bill one month when the board “invited†us to buy up our stock options. Having started with an initial $1m loan from our original investor, Durlacher, we sought our first round of venture capital funding in the summer of 2000. In this first round we raised $8.8m and we had two further rounds in October 2001 ($14.5m) and August 2003 ($17m) for a total investment of around $42m. Our investors were a mix of European and American venture capitalists (including Lehmans, Cazenove, Index ventures and Mosaic), which was a deliberate policy as we guessed that our likely future lay in American ownership or at least further investment.
Somewhat prosaically, one of the main underpinnings of KVS’s solid growth was “processâ€. A major advantage of the fact that we had initially developed the product within a large company rather than starting up with “two guys in a garage†was that we had taken a very formal process-oriented approach to software development and this approach was retained at KVS. For the same reason we also started with fully formed quality assurance and support functions who were a part of the development process. Yes, it would often frustrate sales people (who were probably thinking “I could knock that feature up over the weekendâ€) but the engineering process did produce a reliable product built on solid foundations. Equally, and probably to the surprise of the engineers, the sales organisation also started from Day One with a ready built formal and quantifiable sales process. This produced the required sales results and allowed accurate insight into sales progress and pipeline forecasting, even if the idea was alien to some of our more red-blooded salesmen. In September 2003 we appeared in the Sunday Times “Tech Track 100â€, which was their listing of the top 100 fastest growing technology companies in the UK. We just missed out on the medals by coming 4thwith a year-on-year growth of 238.91%.

Fig.7: 28th September 2003 Sunday Times Cover
In September 2003 we appeared in the Sunday Times “Tech Track 100â€, which was their listing of the top 100 fastest growing technology companies in the UK. We just missed out on the medals by coming 4thwith a year-on-year growth of 238.91%.
Market growth
In retrospect, as good an indicator as any of our growth over KVS’s five years is a series of photographs that I have of our stands at various Microsoft Exchange Conference (MEC) shows. We started at Nice in the fall of 2000 where we taped our display boards to the supports of our minuscule stand and then picked them up and re-taped them the next day after they had all fallen off overnight. Our hospitality suite (and hotel for some of us) was a very nice old wooden boat hired to us at a discount by one of our investors and moored in the harbour.