Failure
Fail often. That’s what the startup people say. They can make failure sound almost joyful. And it’s just one of dozens of clichés about failure which we are constantly fed.
As a citizen of the European Union, I often hear how Europe is a society where it’s not okay to fail, that the American culture is more accepting of failure, and that’s why the Americans are so much more competitive. Well, maybe that was true of an older generation of Europeans, but I don’t really see the fear of failure in the culture of my generation.
I’ve heard that until you’ve failed grandly you’re not ready to run a business. Well, surely that applies to some people, but I don’t know that it necessarily applies to me.
To me that sounds more akin to: It’s okay to take other people’s money and crash and burn dramatically until one day you finally create the product or service that will make you as rich as Elon Musk.
All this advice is so black and white and nicely presented that it’s tempting to believe it.
The one thing I do know for sure is this: I don’t want anybody else setting the criteria for my failure. I don’t mind failing at all. I just want to be the one who calls it a failure. And what I call a failure just might surprise you.
What if the startup community has the whole idea of failure wrong? What if failure isn’t a dramatic crash and burn? What if it has absolutely nothing to do with some externally prescribed criteria?
Let me tell you about my nine failures over the past ten years.
AMS: inadvertently coming full circle
AMS was the first product we developed, our first attempt to move the consultancy into the product business. It was a ‘central station automation system’, collecting and displaying events and alarms from security companies’ clients homes and offices.
It was 1995, and I was able to sell it to every single Estonian security company with the exception of the biggest one. At one point, the majority of the ten local security companies used it. It even provided quite meaningful revenues, totalling a bit more than 25'000USD.
Why’d it fail? First, at the time I didn’t have the expertise to take it abroad, so once I’d sold to my own small market (ten companies!) that was as far as it was going to go.
Second, even though the product could be sold for a couple thousand euros, it required one full day per week of training the security guys on customizations. I could (and did) charge for this training, but this made me feel like I was right back in the consultancy business. What good is a product if you have to spend the rest of your life servicing it?
AMS might have worked out as a business if all I was interested in was money, but it wasn’t the business I wanted to be in. I needed to go back to the drawing board and come up with a product that could be sold without hundreds of hours spent doing service after the sale. I had to look for something simpler.
Digix: getting in over your head
The early 2000s was before the era of cheap Chinese IT cameras which could be connected via WiFi to monitor security. We figured we could do a software for digital video surveillance system, with fancy features like movement detection and pattern recognition.
Our failure was engineering — the product was simply too complicated to create. We built the prototype but never launched it. We actually had to do more consultancy work to pay the bills.
Mobile PDA salesman: killed by smartphones
In the early 2000s one of our consultancy clients was a wholesale company whose employees had to physically check supermarket aisles in order to know what to re-stock.
We created new software for existing Compaq commodity hardware — a PDA similar to Palm Pilot — to serve the one client, and then we sold the application to other companies.
The application was novel at its time. We got some traction with it, and began earning some revenues, but in the end we couldn’t make it scale. This was probably because the app itself was too simplistic, but also the advent of smartphones would kill the niche pretty soon, anyway.
Perhaps just as importantly, we found out that we were not the world’s most aggressive salesmen. It’s one thing to be able to create a product. It’s quite another to have the type of personality that allows you to be relentless in sales.
The mobile PDA wasn’t necessarily a bad idea, but it was an idea that didn’t really fit with our personalities and company culture.
Toggl: the one that never went away
Toggl is certainly not a failure, but I list it in this chapter because it’s important to view it in context. It was fourth in our chain of efforts.
We launched Toggl in 2006 and were encouraged by a few thousand sign-ups in the first two months of business, yet unlike our previous projects it did not generate any revenues.
We couldn’t stop looking for new projects that might generate revenue, but we also couldn’t abandon our several thousand new users. So we decided to keep Toggl around. Engineer readers may have already surmised that I like serial processing, and so Toggl might be characterized as the one product that never went away.
Smartfid: Too early on the market?
Smartfid (pronounced smartfeed) was a 2006 project with Tallinn Technical University scientists exploring RFID cards. The idea was to be able to control a fleet of large vehicles in an easy way.
The Tallinn bus station had hundreds of busses coming and going through a zone that, for a variety of reasons, was not yet physically closed. There was the problem of people parking cars where busses were supposed to be.
The classic solution of wooden barriers activated by drivers reaching out the window for card contact with a sensor was too expensive when it came to hundreds of busses. An RFID system would also mean that they wouldn’t have to stop. It would be very similar to the EasyPass system on US highways, but multiple times cheaper, as low as 50 cents per bus to equip them. It was beautiful.
We built this product, sold it to them, and they’re still using it. Then we sold it to the region’s largest bakery. Then we got stuck. Maybe we sucked at selling — in our defense the B2B sales model was too labor intensive and not our thing. Maybe we were too early on the market with the technology. Whatever the reason, this product ended after two customers.
Hockeystick: serving your own needs
Hockeystick, named for the shape of the curve made famous by climatologist Jerry Mahlman, was an analytics tool to track the lifecycle of Toggl users.
Google Analytics was too general for us, and we needed something focused on cloud startups. A customer signs up, becomes active, starts paying, and so on — that’s the lifecycle. For Toggl, we need to know where in that cycle is the most likely to lose a customer so that we can automate the process to arrest his departure.
We built this product, but the technical challenges were too much for our resources, which is a nice way to say we didn’t have enough people. Plus we had to work on Toggl, plus all our other consultancy clients.
I still believe there’s a huge market for a tool like this. Maybe we’ll come back to this idea at some point?
JetSale: the twitching half of the snake
In 2008 the recession hit. Toggl wasn’t generating revenue yet, our consultancy had slowed down, and we needed money to pay the bills. We decided to systematically examine all the things we had done for our customers to see if there was anything we could turn into a product.
We hit on a simple plane ticketing application we had developed for a tiny airline that serviced a few Estonian islands.
There weren’t any off-the-shelf ticketing systems around at the time that were good for small airlines. Amadeus is the big system, complicated and clunky, and suitable only if you’re British Air or Lufthansa. But if you had five routes and ten planes, then your needs were completely different. If a small airline doesn’t have its own direct sales channel for tickets then it’s going to give away huge sums to brokers. We had developed a simple system that little airlines could handle themselves.
We pitched it to some small aviation companies. We had a Sicilian airline interested but they disappeared after about a half dozen emails. We didn’t make a single sale. For some reason the product/market fit was not right.
As I look back at the recession and how we went about things, I think of a snake. If you cut a snake in half it will twitch. We were desperate. I will give us credit, though, to be smart enough to focus on the twitching half. But a recession is hard to beat, and by late 2009 we’d gone from 20 employees to about ten. But we were still in the game.
8Lanes: great products nobody wants
8Lanes was a spinoff of some work we did for Estonia’s National Athletics Federation. This was a full-featured tool for running track and field competitions. All the disciplines — running, jumps, throwing, combined events, etc. Toolkits for referees, TV commentators, all the cool stuff. We thought this product was really good and were sure everybody in Europe would want to buy it.
We sent personal letters to every general secretary of every athletic federation we could find. Nobody answered. Then we tried direct mail to every athletic club in the UK. One club responded asking for a price quote. We bid 2,000 euros. Their response: “Are you kidding me?” We finally realized that federations of all sizes have limited budgets, and there’s virtually no money in track and field, except at the highest levels. Even the big companies like Seiko were basically doing charity work with their sponsorships.
Maybe it could have been a success. The product is still used to carry the Estonian championships and has even been used for a recent European Junior Championships meet which was held in Tallinn. In that respect it’s a proven product, but we couldn’t manage to get a single customer outside our own country to even test it. Was this a failure? Whatever it was, it was weird.
MySpedition: learning our lightweight approach
In 2010 my good friend Tarvi, who owns a trucking company, told me about a cool product someone had made which could calculate the cost of freight between two points. If you wanted to send two tons of pillows to Germany, it could quickly tell you the approximate cost.
Our idea was to create a simple landing page with simple calculator for German and Polish truck companies to calculate costs.
Trucking is a capital-intensive business. A single truck tire can cost over 1,000 euros, and they’re very influenced by the price of oil. And although trucking companies have small margins, they’re operating with a lot of cash. We thought that if we could produce software to make them more efficient, then even the conservative Germans would buy it!
We were a bit wiser at this point. We knew not to spend too much time producing marketing materials, but to just get the product out there and get people using it. Ours was a lightweight approach with low risk. We quickly created a logo and landing page and put them out there with our calculator. We spammed hundreds of trucking companies to let them know this simple product was now available for them to use.
The result? Zero. And though I have theories, I have no real idea why. We had a simple test product, yet it failed to get any meaningful traction.
I learned a couple of things about business and myself from the trucking experience. First, I learned that if you don’t really know a business inside and out, then this will show in the first few sentences of any written communication. It’s so obvious who’s an insider and who’s a bullshitter. I bet it’s true in your business, as well.
Second, I finally recognized and admitted to myself that I should only work on things that I really believe in. Setting up a product is a time consuming process. The trucking product took us six months to execute with one programmer spending a half-day on it every Friday. How could it have been successful with a half-hearted commitment like that? If we’d believed in it we would have at least committed one person full time for three or four months.
Meanwhile back at the ranch
The good news is that what we’d learned about lightweight approaches was helping us with Toggl. While we were messing around trying to revolutionize the trucking industry, Toggl was growing bit by bit.
The product was earning a little revenue from user subscriptions, a few thousand euros per month, and we were seeing slow but steady growth (with a bit less than 10'000 daily active users). But we still had no way of knowing whether it’d be successful or not. Toggl was in limbo: too good to kill, but not good enough to make it our focus.
So we continued to try new things.
Freshart: another business I knew nothing about
It was 2011. A friend of mine ran a print shop which specialized in making prints on canvas (giclées) from paintings by popular Estonian artists.
I saw that the art market is broken when it comes to regular consumers who just want something nice to hang on the wall. If you weren’t hanging around in art circles, about your only option was to go to the home decoration store and buy some crappy, framed “art” and put it in your home.
Our idea was to take young artists with no access to the real art market, create high quality giclées, and offer these to regular people. The cost of the print was low enough that a giclée could be sold for 200 to 300 euros and we’d still make a profit.
For these “regular” people, non-experts in the art world, buying art is about trust: Can I trust that the art I buy won’t be laughed at by people who come to my home? We wanted to give people assurance that this was cool stuff. “Official hot new artist from London” in limited editions, with new artists added every three months. Think of it as the Zara of the art market, selling quality printed copies of modern art over the internet.
I quickly realized that making a really good replica of a painting is not that easy. It required tech skills that we just didn’t have. And I couldn’t find a local expert who could step in and take over. Like MySpedition, I’d stepped into another business I knew nothing about. We decided to drop the project.
I was on the verge of an almost-epiphany: these products weren’t only products, they were therapy for me to fill some kind of void. Yet, about this time, my partner Krister arrived from his holiday in Turkey with the idea for us to make a new product called Planner, which we would soon re-name Teamweek.
Teamweek: because I couldn’t say no
By 2011 the recession was gone and Toggl was going well enough that we’d dropped all our consultancy projects. But suddenly — and I have no certain, rational explanation for you here — I agreed we needed another product.
Teamweek was that product, an online project planning software with a team calendar. We like to think of it as a kinder, gentler version of Gantt charts.
My best theory of my behavior is that having a consultancy makes for intense days. The pace is quick and you are never at a loss for something to do. Once you switch to a product environment, you have to recalibrate to a new way of working. Having a new product is one way to delay the recalibration.
In 2010, our consultancy client list consisted of 33 clients. By mid 2011, we had zero clients because of Toggl. If you’re used to certain levels of stress, then the lack of it is a void you need to somehow fill. I was a productholic.
I’m happy to report that I was right to go along with Krister. Teamweek has now generated more revenue in absolute terms than all the other “failures” combined, and it has generated more active users by a factor of 100. So what’s different with Teamweek? What did we do right?
For one, we applied a lot of what we learned from Toggl. Teamweek has similar business attributes to Toggl — a monthly fee and it’s a product offered over the internet as a cloud application. Since this chapter is about failures, I should add that we thought we could cross sell it, but that didn’t work.
Another thing we did right with Teamweek was not to try to hedge our risks. We put two developers on it full time. We released it in a couple of months without ever a thought of dropping it. Now it’s part of our portfolio.
Expertise: finally getting it right
Teamweek also fell into a category of product that we actually knew something about. In art and in trucking I had friends who possessed expertise, but none of them were ready to become co-founders of a startup. You can’t go it alone, and knowing how to build software just isn’t enough.
But time tracking was my area of expertise. I’d lived the consultancy life for ten years myself and knew the problems of team organization and planning inside and out.
When I think back on my days as an expert on trucking, I imagine what a good laugh my pitch letters must have given to the Germans in the trucking business. There is a language and a dance specific to every industry. The German trucker surely smelled me coming before he even opened the letter. It’s just so very hard to bullshit in B2B sales.
Why we failed
The majority of products on my failure list will probably not meet the traditional definition of failure. Nothing went bankrupt. No external party forced us to drop them. In all nine cases we chose to abandon them. A friend of mine once commented that they are some of the most successful failures he’s seen.
So why’d we kill them? Why do I consider them failures? Many had no visible light at the end of the tunnel. About as soon as we’d built them we could tell they weren’t going to work out. There are many potential points for failure, but it’s sometimes hard to see them before you actually start building the product.
Sometimes we failed because the sales cycle appeared slow or complex. In a few cases we didn’t set up a proper team. Other times our product was too early, or the market was too conservative, or an engineering task turned out to be too complex for us, or we didn’t put in enough effort to make the product scale.
Often, the amount of time consumed and hassle received was just not worth the effort we expended. They didn’t make us happy.
Sometimes we just couldn’t sell. But I can’t fault myself too much for this. To sell you need more than a salesman. You need a system to sell, an entire plan with activities and follow up. You need to cover all the trade shows. Selling is something that we just weren’t internally wired for, and consequently our organization wasn’t built with sales as a priority. But, of course, sales cannot be avoided. You just have to do it your way.
Our initial approach with Toggl was different. We didn’t start with a product. We started with a market. We thought: Who do we want to sell to? And how do we want to sell to them?
The conventional wisdom about failure
In the US, there are lots of dramatic stories about failure. You hear of suicides over startups. You hear of people amassing tens of millions in venture capital and then burning it all. My failures are much more boring. Luckily, you can learn from a failure even if it isn’t epic.
If there’s one thing I learned, and pardon this cliché, it’s to trust your gut. And if you don’t trust your own gut, remember that you can also check the gut feelings of others on the team.
When I went to the team I generally found that my personal feeling was also the general feeling of the team. And just from observing the team, I could pretty easily see what they were excited about. Sometimes it wasn’t that one idea was terrible, but simply that another was better. Some products felt like work; others felt like play.
In most cases, however, the rational won’t let you down. If you can detach yourself for a moment, you can see what assumptions are flawed. In a consultancy you’re constantly approached by customers with dumb ideas. Everybody gets it wrong once in a while, but in the vast majority of cases you know which of these companies will be quickly insolvent. If you can see it in others, then you can learn to see it in yourself.
Maybe that’s the real benefit of failures: they help you see yourself more clearly.
I submit that the conventional definition of failure is wrong. There’s a slogan on the wall of a startup bootcamp: “You’ve got what it takes, but it will take everything you’ve got.” Business coaches and VCs are fond of saying that startups are hard. They talk about the death zone of Mt. Everest. They say whatever you do you’ll be depressed, you’ll run out of money, you’ll have no family life.
I say a startup can be a wonderful walk on the base camp trail. You can hike a long time without getting hurt. It doesn’t have to be something that’s unnatural. I don’t see a death zone. I see an enjoyable experience that you can do even when you’re an old man.
I think what’s important to keep in mind about failure is that it’s a failure when you say it is, not when external criteria say so.
It doesn’t have to end in tears.
