I recently announced my decision to leave FreeAgent, the company I co-founded in 2007. It's been a difficult decision to make as I love FreeAgent dearly – I’m part of the furniture almost – but it’s the right time for me to hang up my well-worn CTO boots. I wanted to write a few words for posterity, so here are some of my memories from our rollercoaster journey.
5,387 days ago (19th Feb 2007), having recently quit a contract where I was completely downhearted working on an unfathomable credit derivatives product at RBS on Bishopsgate in London, and with a bookkeeping web app called One Man Band running through my mind, I contacted Ed after seeing a message he’d posted on the back channel of the Future of Web Apps conference which we were both attending. The message said something along the lines of “Come and talk to me about FreeAgent Central: Solo Simplified” and there was a landing page at freeagentcentral.co.uk with hot air balloons. I was both deflated (unlike the balloons) – that was my idea dammit – yet excited, because Ed had actually produced something that looked great, was written in Ruby on Rails, and it sounded like there might be a chance to get involved.
We met in the basement of a conference centre in South Kensington and I had a look at the “FreeAgent Central” prototype which looked pretty good. It had animated sliding help menus (RIP) and everything. I was impressed! We had a good chat, I said I was up for getting involved, and that was that. What followed was a blast.
I’d done some interesting work before: launched games, PC desktop publishing products, internet radio stations. I was already working for myself since I quit my full time job in 2005, set up a limited company and got a 6-month contract in a bank. So there I was, wearing a shirt, trousers and polished shoes, getting sardined on the Northern Line to Moorgate each morning but at least I was sort of a bit more in charge of my own destiny.
My long term plan at that point, or rather what was in the back of my mind (I had no plan), wasn’t to be a full-time contractor for the next 20 years (a sensible and lucrative option which was encouraged by my accountant at the time). The idea was to milk a couple of City contracts for as much money as possible, live reasonably frugally and build up some savings in order to try and make a living actually working for myself on my own product. Inspired by “Web 2.0” companies such as Basecamp, Flickr, De.licio.us, this lifestyle felt in reach. I did some client web design on the side but quickly discovered that this was hard work. It was frustrating dealing with nitpicking clients and the general return on investment was poor. I tried the software agency model of pitching to businesses to build web apps, but I was terrible at it.
Meeting Ed (and subsequently Roan) changed this. FreeAgent was solving a problem that I wanted to try and solve, that I actually had a bunch of experience with, using technology that I was eager to learn. It was a classic case of right place, right time.
I was lucky enough to have a decent amount of savings in my limited company so, newly liberated, I started working on FreeAgent every day from our flat in Kennington (ok, Walworth). Ed was still doing some ad hoc consultancy and Roan was still full time in the day job and working evenings at this point, so we used Basecamp to plan work with the occasional Skype call to make things feel a bit more real.
At this time the app had some core functionality and it was well-tested (following The Rails Way). It was my job to add a bunch of missing functionality, polish things and get the product into public beta and ultimately launched, which we planned for later that summer. I loved it! As well as coding and sorting out basic infrastructure (this was ‘pre-cloud’), I started networking (friends, acquaintances, local freelancers from meetups I attended) to talk up the product and to get early feedback. There was definitely interest, but general excitement levels were far lower than my own. We released features weekly (we hadn’t yet learned the lessons of continuous delivery!) accompanied by cheese-themed blog posts. During the summer we came to an arrangement to allow me to keep working full time until the end of the year, with Ed and Roan continuing with the day jobs.
By the time September rolled around, we officially launched 🚀
Build it and they will come 🤣
The product came with a 30-day free trial which gave me four weeks from launch day to implement a payment system (it literally took this long – now you can do it in approximately four minutes using Stripe) that would allow the $millions to start flowing into our company bank account so we could get the Ferrari’s on order. Bring it!
In the first week we got 1 paying customer. £24.
Over the next month or two we started to see a few dozen customers arrive, but it was pretty clear we wouldn’t be setting up payroll any time soon. We had zero budget. By this time all three of us had new babies arrive at home, so with no money coming in I was forced to go back to my former employer who kindly offered me a 6-month contract (I think it might have been 3 days/week). We managed to keep on improving the product but it was definitely slower going. Ed focused on sales with some dev on the side, Roan and I focused on developing more features for the app. All of us handled customer support, networking, tweeting, spreading the word. Building a company is, for the majority of entrepreneurs, a hard slog and a long, grinding road. It certainly was for us, but we weren’t being ground down – we were still really up for it and believed in the mission.
Wallets to the rescue 💸
Towards the end of 2008, Christoph Janz got in touch with Ed after seeing us mentioned in a TechCrunch article we probably badgered Mike Butcher to write. Christoph is a legend and he introduced us to another legend, Robin Klein. Between them, we secured our first investment of around £200k along with another £50-100k or so from friends and smaller angels. This is not a lot by today’s standards (I currently see pre-seed - never mind seed - investments of £800k+ 🤯) but enough for us to accelerate things. Doing more with less has been a recurring theme for FreeAgent, and we got pretty good at it.
We hired a sales person on contract, focused on accountancy practices as a new channel and started to see some early success. We had about 1,500 customers in total. Ed was busy pitching to investors and an early lead with a Scottish VC led to an introduction to IRIS, the UK accounting software goliaths. IRIS ended up investing £800k with an exclusive (for a while) deal to sell FreeAgent (co-branded as IRIS OpenBooks) to their accountancy practice clients. We finally had some cash and could start paying ourselves more than £500/mo. It only took about 3 years. We were off!
Moving country 🏴
Although we’d been a remote company for 3 years, the plan was always to get an office in Edinburgh whenever we had the funds to do so. This was a big deal for me and my family as we’d been in London for a decade and had friends and family there. Edinburgh seemed a long way away (it is) but the opportunity to get an office and hire staff was a big one, and I felt that I needed to be there. Things were getting real. Looking back, we probably could have stayed in London and I could have commuted, but that would have been detrimental to family life and, frankly, London life in a small 2-bed with two children was getting tricker and we couldn’t afford to move somewhere bigger. There was also a murder a few yards from our flat towards the end of 2009, so getting out of Walworth to go, well, anywhere, seemed like a pretty good idea. Edinburgh it was!
Our first official office was on Miller Crescent in Morningside which we’d rented for a few months in 2009 (being in London, I only popped in once). Following the IRIS investment, we upgraded and moved across the road to a shared, ant-filled former prison/office on Canaan Lane where we lived for the next couple of years, gradually taking it over and expanding into the neighbouring office (the “dev pit”) when our engineering team got to about 7 or 8 people. They were happy days!
By 2011 we had about 6,000 customers and growth was being fuelled by a deal which saw FreeAgent (and a few other SaaS products, and ultimately Sage and QuickBooks who I think felt left out) in a package offered to new Barclays Bank business customers. We took more investment from SM Trust in 2011 and moved to our first proper office, complete with fancy fit out, on Torphichen Street. I loved that office! The company was around 30 people by this stage, the IRIS deal was going reasonably well and we started to establish ourselves as one of the main players in the UK. I remember fundraising (for Ed) being a constant battle. We raised money, hired people, added features, grew the customer base… rinse and repeat. We were never profitable, it was always about the next runway.
The next fundraising effort was probably the oddest. The founders of Groupon – Brad Keywell and Eric Lefkofsky – had a startup accelerator called Lightbank in Chicago. They were interested in FreeAgent and there was an opportunity for them to invest which we felt could open doors to getting FreeAgent more established in the USA. We had already made an international version (“FreeAgent Universal”) which had a bit of traction despite zero marketing effort, so imagine what we could achieve with a farm of salespeople in a Groupon call centre bringing in leads 24 hours a day? Caribbean islands and private jets ahoy!
This is what was in my naive head at the time, so we flew to Chicago for 36 hours, barely sleeping, and I forgot to buy travel insurance so I had to live with the fear of my general clumsiness resulting in bankruptcy-inducing hospital bills. I got away with it.
The Lightbank deal on the table involved a modest cash investment but also involved acquiring a company, 60mo – a financial forecasting app. To be fair last thing we needed was to acquire a company that used different tech and was based in a city 3700 miles away, but we went ahead with it anyway. We were young and needed the money, but we felt getting Lightbank on board would allow us to kickstart our presence in the US (we even set up FreeAgent, Inc. – a Delaware corporation – around this time). Lightbank obviously wanted an exit for their investment in 60mo and the FreeAgent deal was a tidy way for them to achieve this. We made the most of it – the 60mo founders came on board and helped us build a specific US version of FreeAgent as well as our first mobile web offering, but we ended up parting ways a couple of years later. I felt terrible about this, still do, and it was definitely one of the low points.
The following years took a similar path with customer growth being good, the product constantly getting better and investment being a constant battle. In 2013, with customer numbers over 30,000, we received an acquisition offer which went down to the wire but failed because we demanded (purposefully, sort of) too much. It was pretty exciting and something of an adrenaline rush that made the daily grind of hiring, building features, arguments about prioritisation, people issues and so on disappear for a brief period. Looking back, that acquisition would have been a better outcome financially but it was the right decision at the time. I remember being really up for it, not just because of a cash bonanza but I thought there was potential for a larger impact. At some level (like most businesses) we always struggled with not having enough resource (cash, people, time) and this deal had the potential to change all that, but on the flip side it could equally have been a disaster. Things happen (or don’t) for a reason.
Hot on the heels of M&A term sheet drama, we had another potentially big private equity investment which was equally exciting but ultimately didn’t work out, so we ended up taking on convertible debt, then some venture debt. Amazingly there were still some fundraising techniques we hadn’t yet tried, so in 2015 we ran a crowdfunding campaign which was a total slog, but we got there in the end and raised £1m from hundreds of FreeAgent customers. Thank you! 🙏 We moved to a larger office – our amazing current abode – and no doubt starting thinking about the next fundraising round.
A glimpse of the future
In 2016 we took part in a pitch to RBS/NatWest for a co-branded pilot they wanted to run with an accounting platform. We were thrilled to be selected as their preferred product. We had 50,000 customers and started working with the bank to get some of their customers onboard, and to develop direct automated bank feeds to make the integration feel really slick (automated bank feeds were still a problem at this stage, pre-Open Banking).
At the same time something else was taking shape – we had planned to take the company public on AIM. This was the largest (and most expensive) fundraising we’d done. We’d see great PR, a chunky cash injection and could (in theory) raise capital more easily in the future. An IPO would also be the ‘liquidity event’ that would allow our early investors, as well as our longer-standing staff who had options, to finally cash out (no such luck for the founders though as it turned out secondaries are frowned upon in the City). Our numbers looked pretty good and the merchant bankers were happy! An enormous amount of work followed over many months, mainly by Ed and Kath, our CFO at the time – I just had to deal with the tech due diligence which I’d been subjected to several times before.
On 16 Nov 2016 we ‘rang the bell’ (pressed a fake button) in the grand atrium of the London Stock Exchange. Y’all were free to buy shares in FREE! 84p each. Bargain!
We knew it could be tough on the public markets. If you consistently met your projections (and didn’t sandbag) you would be rewarded with a buoyant share price. If you failed, you’d be punished and see the share price crash through the floor. This seems pretty brutal because it is. We had the common ‘honeymoon’ period where the share price grew 50% or more, but after a while it was hovering several points below our IPO price despite growth being pretty good and our financial projections being fairly spot on 🤷. It definitely had its moments, but I found this period a little tough because the vibe (for me as an exec on the team – hopefully not for most of the staff) changed. It kinda had to as we were now a PLC. At the end of a really exciting chapter, the start of the next one is always going to be a little underwhelming. Hard act to follow.
In early 2018, with the share price down but the NatWest arrangement working well, we received a very unexpected term sheet – the bank wanted to acquire the company. And after a lot of negotiations, they did! Funny how things turn out.
It seems pretty obvious in hindsight, but we genuinely had no idea that an acquisition would be ‘a thing’ at the time, but I’m glad it happened and not just for ££ reasons. It was, without question, the best outcome for the company since it brought investment, stability and a heap of opportunity. The last three years have validated this, especially during the pandemic. We’ve grown the business faster than ever, customers still love the product and there’s a rosy outlook.
It would have been tough to do that alone.
I have no plans for what’s next, unless being focused on trying not to do anything in particular is considered a plan. I’m tired and I need a break – 15 years is a long time to be on call.
Starting another business, at least a high growth venture-backed startup, sounds a bit tiring right now but I wouldn’t rule out an indie hack – I still love to create. I’ll see if there’s interest from other companies in advisory or non-exec positions and I have a list of “articles to write” which I’ve struggled to make a dent in, so maybe I’ll write more. And take more photos. Or maybe I won’t!
Time will tell.
I’ll sign off with a few stats from my time at the company. I have no doubt these numbers will keep on rising, and I’ll be stoked to hear how things evolve in the years ahead.
E praeterito ad futurum! 💚💙
- 120,000 Customers (~690,000 accounts created)
- 260 Staff
- 1 Edinburgh HQ
- 70 WeeAgents (babies born since we founded the company)
- 1 staff marriage
- 34,000 pull requests
- 142,000 commits
- 250,000 lines of code