Do one thing and do it well

20 November 2008

It’s all change at Resolver Systems.

When we founded the company at the end of 2005, Robert, Patrick and I wanted to produce a new kind of spreadsheet. It would be something that would be bought by individuals within a company, using their company credit cards, and would gain market share slowly but surely, by making its owners more productive than their peers.

Based on this plan, we raised money, started putting together the team, and worked hard to get a good first cut of the software ready for our first milestone — a meeting of the advisory board.

At that meeting our advisory board looked at what we had, looked at the spreadsheet market (with its dominance by Microsoft Excel), and told us that our plan couldn’t work. The real money in the spreadsheet market was in the big-ticket sales, they explained. Sure, we could sell a souped-up spreadsheet for programmers and make a few bucks, but it we could produce something… I hesitate to say enterprise-grade, but that was the implication — well, then we could start looking at making real money.

We weren’t entirely convinced, but these were older and wiser people than us; people who had made their fortunes in the (enterprise) software world. So we tried to work out how to best do this. There were various iterations of the plan, but the aim was always the same: to produce a high-end version with a high price tag for the top end of the market, in particular for the financial markets, and a another version for everyone else.

It took a while, with constant adjustments of what should go into each version, but this summer we finally cracked it. On our website we would offer a retail version of the software (and a free version for not-for-profit use), but the bulk of our marketing, and all of our in-person sales effort, would go into identifying large finance companies with specific spreadsheet problems, for whom we would provide services around our core platform as a way of giving them a high-end product — with a high-end price tag. In my wilder moments, I found myself thinking that we could open-source the platform and get our revenues entirely from these services, or from tools built around the platform — like 37signals have with Ruby on Rails.

What could go wrong? After all, our only dependency was on the market for highly priced software for financial companies.

Right.

In October, our top-end customers started cutting back their requirements for new copies of the software, and at our weekly sales meetings, the “potential blocking factors” column in the status spreadsheet started sprouting worrying new entries like: “Client being acquired by a non-bankrupt competitor, all expenditure on hold. End user still keen.”

It was clearly time to re-think.

I don’t know who it was that said it, but I remember reading once advice saying something like: “Imagine what you would do if your company was in real trouble. What would you do to bring the company back from the brink? Now, for each of those things you would do then — why aren’t you doing them now?”

To put it another way, the prospect of impending demise sharpens the mind wonderfully. We’ve come to realise that since that original meeting, we’ve been trying to be two companies; an enterprise services vendor, selling software based on our own platform, and a vendor of a new kind of spreadsheet. We’ve been putting our management and sales effort into the former, but we have the DNA of the latter.

(At a developer interview we’re much more likely to ask someone how they might go about writing an algorithm to topologically sort a dependency graph than how they might calculate the value of a CDS. That’s a free hint to anyone thinking of applying for a job.)

Our problem wasn’t that we’d followed our advisory board’s advice and they were wrong (though it would be easy for us to see it that way). If we’d followed their advice entirely and given up our dreams of producing a new kind of spreadsheet, we could probably have produced a fairly decent humdrum kind of financial software services company, which would be in trouble now, but with good management might have been able to weather the storm. It wouldn’t have been the company we planned for, and we probably would have had to be replaced as a management team at some point, but it could have worked as a company. If, on the other hand, we’d rejected their advice outright and gone ahead with the regular commercial product, with appropriate marketing, we could have been a significantly bigger force in the spreadsheet market by now — as one of our users put it in an email, we could have got further in “rescuing the world from VBA” (a worthy cause if ever there was one).

The net result of this is that we’re simplifying; we’re going whole-heartedly for our original plan, and dropping the enterprisey nonsense; like a Unix shell command, we’ll do one thing, and do it well. Naturally, we’ll keep to our existing commitments to our enterprise clients, but from this release, Resolver Systems is a company producing a spreadsheet product. It’s a souped up spreadsheet for people who don’t want to make costly mistakes. For people who want to be able to see and control the steps taking place as part of the spreadsheet’s calculations. For people who want to know that if they add a new row or column, they don’t need to remember to change umpteen formulae elsewhere.

It’s called Resolver One, and we released version 1.3 this week. Let’s see if we can rescue the world from VBA.

[UPDATE] Coincidentally, Dharmesh Shah at OnStartups just made a post arguing for precisely the opposite strategy :-)

[UPDATE] Elsewhere, Laeeth suggested that the “do now what you would do if you were in trouble” might have been Andy Grove of Intel, who said something like “if we were fired and replaced by new management, what would be the first changes to strategy they would make?”

4 thoughts on “Do one thing and do it well

  1. giles Post author

    Thanks, Corey! We’re feeling pretty hopeful, the 1.3 release has gone well and psychologically it just feels like we’ve got stuff moving again.

  2. StefanTh

    Sounds fantastic!!!

    Next step would be a competition where people build models from the real world of spreadsheet hell, book examples would be from advance structured finance vba books, and such, that exist on the market.

  3. giles Post author

    @StephanTh – absolutely! We’re working on an addition to our website, provisionally named “the Resolver Exchange”, where people can upload any Resolver One spreadsheets they want to share with other people. A competition would be a great way to encourage everyone to upload as much as possible and make it a really good resource :-)

Comments are closed.