In 2004 — wow, I’m just realising that this is exactly 10 years ago now :) — we have started a hosted forum service — a wordpress.com for forums — under the name of Ultraboard (later renamed to Parlaris) where you could create a free forum on a subdomain, e.g. yourforum.ultraboard.ch.
When deciding to build this service, we set ourselves two options: build a hosted blogging service (like blogger.com) or to play it safe and focus on good ol’ forums. Blogs were sexy and trendy, but not yet mainstream. At the same time we were no bloggers ourselves, but were heavy forum users. So we decided to play it safe and chose the forum option. This was probably mistake number one, because it’s probably more fun and more rewarding to work on something trendy and sexy than something old fashioned. But it’s of course easier to say that in hindsight.
Mistake number two was, that we focused our business model on advertising. Premium subscriptions were no option for us, so we have built the product without any premium upgrades in mind.
After the first year we learned that our assumption of $3 to $5 eCPM was way too high. The reality was more like $0.30 to $0.50 eCPM which made our business model very weak. We then desperately offered a premium subscription with the main value proposition of being ad-free plus many other things packed into the premium option, actually the one and only premium option — we offered no different options. This was mistake number three.
We failed to communicate the premium option well enough and failed to offer a compelling enough incentive to upgrade. So after the first 12 months after launching our premium service we only had the great number of 14 subscribers, of which 8 never paid (as also we failed to integrate an automated payment service like PayPal). This was mistake number four.
We had more than 5000 forums with more than 100.000 users, with millions of pageviews per month and more than 55.000 hours of usage time per month (= roughly 430.000 visits per month with an average spent time of 8 minutes per visit). So it wasn’t that the service was not successful. Users found value in it and used it. The problem was that we executed very badly on the business side. We only made roughly $650 gross margin per month (top-line revenue was $1000 per month). Considered that we’ve spent roughly 600 hours building it and many hours per week for support and maintenance, this was not sustainable by far. Therefore product development and marketing expenditures made no sense with these kinds of numbers.
The advertising revenue was lousy and the premium business was even lousier. Instead of increasing the incentive to upgrade, we were simply frustrated and didn’t change anymore. This was the final mistake. Then the numbers began to fall and our frustration grew. In the end we sold the service in 2009 — after struggling for more than 5 years — for a fairly small amount to a webhosting company. Unfortunately the integration into the new company didn’t work out and the service was discontinued in the end.
It’s a pretty sad and frustrating story, but also an experience I have learned from a lot. Maybe you can learn something from it as well. There is nothing more frustrating to run a service and knowing that it is doomed. We were not courageous enough to pull the plug sooner. We should have pushed some more radical changes, set some goals and if these wouldn’t be met then sell or quit. Quitting is not the only option. You can always try to sell your product or service and try to achieve a ‘soft landing’. I believe if we had tried to sell it some years earlier we could have sold it for a much better price because the project had more traction. We also would have saved us some years of frustration. While this project obviously didn’t consume all our time, it was constantly in the back of our head, consuming valuable energy that we could have invested somewhere else (= opportunity costs). And maybe someone would have figured out a better business model and the service would have survived until today. Who knows.
So long, R.I.P. Parlaris.