Yesterday evening in peak time we had serious capacity problems on our UK service. For more than one hour, you had to be very hungry, patient and persistent in order to place an order. We deeply apologize for this very bad service to our customers, restaurants and partners!
The reasons for the problems can be boiled down to one thing: we had a major burst of nation wide UK TV advertisement that evening, and the ad’s did exactly what they had to do: make people curious. Even though most new visitors did not place an order, they did browse and test the system. And they did it in much larger numbers than we had estimated, which meant a positive turned into a negative when servers running parts of our service slowed down to snail speed. You can read more about the tech problems here.
It is not good enough. We should of course be ready to take incoming traffic well above the marketing estimates, so there is no excuse here. This morning a hand full of the Tech guys added and upgraded servers, so this evening – where we again had a lot of TV slots – things went much better. But some customers still had a bad experience. Luckily, we don’t have TV commercials the next couple of weeks, so there is time to improve to the right service level.
Belly cries, because he is hungry here and now, Brain, however – I think – is smiling, because he knows the capacity problem is almost solved, and the fact that TV worked well for us (expensive, yes-yes) means he is going to be very famous in the long term.
To add to our “TV success” challenge is not only the number of servers, but also a system, that has been build up over many years. Often systems and sub-systesm has been added, changed and deleted without the “grand overview”. The solutions might have made sense when they were made, but today, when the company is 10 times bigger, they are too expensive in server time. Therefore, the real solution is not only to add more iron to the server park, but to optimize the code base – significantly. We will fix the problem here and now by adding servers, but when that has happened then we will focus our attention to a smarter way of mitigating growing pains: smarter systems.
There are no news in this for us, and it is a problem most successful internet companies experience, but there are no excuses, and we need as a team to show that we can master this conversion from small start-up to big business also on the Tech side. 9 months ago, our entire Tech team was in Århus, Denmark, and it was only 6 full time employees (plus one Product Manager), which was tiny for an internet company of our size. Then we got our series A, we got Carlos on board as CTO, and we started expanding the team. Today our Tech team incl. Product Management is almost 25 people, and most of them are in London – and we are looking to hire another 8-10 people ASAP, some of the job ads you can see here, e.g. developers. We only hire top people, but if you are one of those, send your CV.
The classic combination of high growth, too small a Tech team from the beginning and a non-optimized and complex system will be defining for the future of Just-Eat. If we have to put most of our Tech power into making sure Just-Eat.co.uk will always be smooth and fast (which is the key priority the coming week, so this does not happen again), then we will not accomplish the many other things that we want to do, e.g. launch in many other countries, build even better web products, streamline internal admin tools to bring down costs, etc. I am confident we will be able to be ambitious on both dimensions, but we will be busy for sure. We want to make sure, we help making Takeaway smarter, and technology is key to this.