Calling Time

9 November 2005

After struggling to establish a market for a new kind of product, London-based Choreology lost its VC backer, put its commercial operations on ice, and switched focus to the free 'open source' software community. It's a cautionary tale, and CEO Alastair Green has plenty of advice for other start-ups. 

Choreology focused on the emerging area of business transaction management, a way of managing transactions that span multiple software applications, such as a credit approval process for a loan. Its product, called Cohesions, was designed around an increasingly popular approach to software design, 'web services'. But after almost three years and $12 million of equity and loans, its VC pulled out at the end of last year.

The company has since released a new version of Cohesions - billed as 'more powerful, business-natural and easy to use than any previous distributed transaction software' - and is now handing the code to the product to the open source community. CEO Alastair Green explains what happened:

Were you too early to market with an idea that was too revolutionary?
The problem we had as a company was that business transaction management (BTM), while being an evolution from the transaction processing software that we have had since the 1960s, was new technology. We were therefore dependent on how the incumbents acted - IBM, Microsoft, Oracle, BEA - and they have been very slow at moving into the space. So the market has been frozen while everyone waits to see what the big players do.

In addition, standards in web services are still emerging. The WS-Transaction standard (which governs the way these kinds of products interact) has only just made it into the standards body OASIS. When it emerges in a couple of years' time there might be more movement.

What lessons did you learn from this?
If there's a way of doing what you are doing slower and for less money, then you should take it. The longer you can make the money last and the less you can rely on optimistic uptake in the market, the better. We had 18-22 staff at our peak when we probably could have done it with seven or eight people and been reasonably effective in achieving an end result.

Where did the idea for Choreology come from in the first place?

All these people had come together who had prior experience in other software companies of transaction processing, and for varying reasons had come across BTM as a problem area. We felt there was a big gap in the market and we had all seen cases where the existing technology was not being widely applied in the way things were done. So the idea had been germinating for a long time.

You have some experience in start-ups, having raised seed funding for Newcastle University spinout Arjuna Solutions, before selling it to Bluestone Software for $13m after 18 months. What's the greatest risk in this kind of company?
It's all about timing. In Arjuna's case, its commercial possibility might never have been realised because as it happened, it was hitting the tail end of a market. From the experience of people who have been through it, the biggest complaint about conventional investors is their unwillingness to put the time into understanding what people are doing. I have sympathy for investors because they've got to go by experience and the level of information they are given. But that is one of the distinguishing factors of a good investor.

VCs are always quick to walk out if they don't get the right feeling from an idea. I think entrepreneurs should do the same. You can always tell if people are really interested in your idea.

Having said that, you raised some $12m with Choreology. We're assuming you found one of the better investors?
We originally put together $1m through our own resources and angel investment, and we assembled a team that included as non-executive chairman Mark Hemsley, who has a long history in IT management in investment banking so he knew the end-user standpoint.

We were then contacted by Atlas Venture, who had been looking for a web services investment and had come across us. In July 2002, they put in $3m in seed investment and $4.5m the following year. Thereafter our history was less conventional as we entered a series of bridging arrangements which took us up to nearly $12m.

Why the bridging arrangements?
When they get into a situation (where sales are not coming in), VCs tend to think the problem is the team. So they hire a VP of sales or a commercial director. Commercial saviours tend to be expensive to hire and expensive to get rid of. So you can easily spend a lot of money.

We take it you were not entirely convinced a commercial director was needed...
It depends on the circumstances. We were convinced it was going to happen - it was just a question of when. We never thought the problem Choreology had was going to be a lack of commercial experience - and perhaps looking back it might have been better to slash back and look out for a limited number of technology-oriented partnerships or marquee deals.

But take a different situation. Let's imagine you've sold 20 customers and you want to expand from there. Then the problem is how you put a sales structure in place and get the product out of the door. There is a technical skill involved in that environment. But that technical skill is not what you need to make your first sales.

VCs think they can help shape the team, but it's like if you have a hammer, every problem is a nail. Perhaps they err on the side of things they are closest to - finances and sales - where they might be better helping in other more subtle ways.

So looking back, were you pursuing a market that wasn't there for too long?
You might say it would have been better to decide it wasn't happening and pull the plug (earlier on) but I don't think they miscalled it. It's hard to admit there's not a market or the market is not developed enough to make a business proposition. But we're not blaming Atlas. They're a sensible, level-headed and highly supportive outfit. They believed in the management team and have said they will back any future venture we embark on.

So what happened next?
During the whole of 2003 it was impossible to sell infrastructure software. Companies had an emphasis on cost cutting and 2003 may as well have never happened as far as we were concerned. In 2004 we got a first realistic flurry of bites.

There are two ways to sell software in our world. You can sell it to end users or license it to other software companies (OEM deals) and it's very likely your early deals will come through the latter route. If you've got something very leading edge and BEA or IBM knows it's something that they're likely to want to get into, then you're more likely to have success that way than through end-user architects.

VCs don't like OEM plays - what they think is that you make two or three sales, then you've sold your technology. But the arrangement can work. If you can get enough back and you're not desperate, then you can build up a steady income stream from a good royalty base. And if you ensure you have a good view of how they're positioning it and are only selling it within an existing product, you can actually get good referenceability.

We came agonisingly close, with a serious deal with a serious company - we were on the verge of a seven-figure OEM deal - but in December 2004, Atlas decided it would not continue to fund the company on a sole basis and pulled out.

And what's happened since?
We have just put out our v3.0 release, which is a huge evolution, and we've decided to open source the product. (So) its afterlife may be more elaborate.

Is there a future for the technology?
We think it will come down to sponsoring up to four people to work on the technology. We've had three paying customers in the history of the company and a number of pilots and collaborative initiatives, so we know the software has a use. But we were too preoccupied with trying to get early revenues rather than getting people actually using the software. Utilisation is the father of referenceability and referenceability is the mother of sales in this situation.

Alastair Green was talking to David Longworth of Webster Buchanan Research