13
Dec 12

Lessons learnt from doing startups

Quick brain-dump of some clear, brutal lessons I have personally learnt doing a few startups over the past five years or so. I’ll maybe add to and unpack these at a later date, but these lessons are hard won.

Lesson #1 if it involves someone else’s copyright content, stay away. It will end in tears & disappointment.

Lesson #2 It can *never* hurt to spend an extra week or two just talking through the whole thing before beginning.

Lesson #3 If Google enters your space, pivot sooner than later. Seriously.

Lesson #4 Operate a single veto over senior hires – either total love or don’t hire that person.

Lesson #5 Get your employment contract and shares/options in writing, in clear language. Stick it on the wall.

Lesson #6 For all the user research in the world, if you wouldn’t use it there’s probably something wrong.

Lesson #7 As unsexy as it is, *start* with powerful email marketing and metrics systems from day one. Not an add on.

Lesson #8 If you are spending a week per month just covering your arse for the board meeting, you’re doing it wrong.

Lesson #9 A simple clear vision to rally behind is *not* optional. A startup tries out many things, but if they all align to the vision, momentum is never lost.

Lesson #10 Don’t put money in if you have talent (code, design, hustling, writing) – there’s loads of money around, not much talent.

Lesson #11 Avoid dilution with good employee share pools, and try not to accumulate massive liquidation preferences – you will lose interest when your potential upside < earning potential as a talented n.


09
Mar 11

Week 22

Now, that was a week. We did it. We shipped. You don’t get to do that entire thing, from first chat to spinning up a dozen EC2 instances to cope with *blush* sudden popularity very often.

Artfinder is now live, go take a look. Go on, now – i’ll still be here in a few minutes.

There’s an official blog too at blog.artfinder.com where I will be periodically writing about design, feature and product issues. I’m going to write this one here first to try and collect my thoughts on our process and strategy.

Obviously we’re trapped in the classic startup dilemma of Time vs Resources vs Quality and I would like to talk about the choices we made and why.

Firstly there’s Time, and you may tell from the notes on this blog, and the gap between week 3 and week 19, that we didn’t move quite as fast as we originally hoped. But, at the end of last year we made a plan and we stuck to our dates* which was essential for us. First mover advantage is, in case, critical.

The second part of that equation is Resources, and it goes without saying we are streamlined to the extreme. On the design side it’s been just me so far, although we have an outstanding designer joining us soon. London has become extremely tight for exceptional design talent right now. Feels bubbly.

So with two parts of the equation fixed rigidly, we had to flex a little on Quality. We’ve made choices about features, implementations and shortcuts that have been difficult, but I completely stand by. It was only 4 weeks ago when we had a beta I could test in front of friends and family, with guidance on missing bits and known problems. Only did the iteration 2 weeks before launch did many of the elements fit together for the first time. Truly lean.

We’re very proud of what we’ve shipped, and the vast majority of positive comments has been very moving, but we’re extremely aware of many problems with it.

As designer and product manager I second this quote by Reid Hoffman, “If you review your first site version and don’t feel embarrassment, you spent too much time on it.” (Incidentally, Reid is an investor in Artfinder.)

How has this manifested itself?

“feature x is unpolished” – I agree, we’ve launched with the absolute minimum in place, the minimum of complex code (especially ajax and javascript visual effects) and the minimum user messaging. Are we going to add these missing layers? Of course, and as soon as possible.

“feature y is unclear” – From the feedback we sought during earlier, private, betas we agonised over some of the placement, naming and function of some features. Did we get it right? Certainly not, and we’ll continue to agonise over them until we do.

“feature z is missing” – The feature list of doom is massive, we picked out the ones we felt essential to convey what the Artfinder concept is about. The ones we did implement, we cut to the bone. For example collections – at launch *extremely* limited (you can see 8 artworks! No delete!) but have plans on how the next, next next, and later iterations expand on this.

“content x is missing or inaccurate” – We know, we’re working on it, the copyright issues we face are considerable but we think by being open, positive and doing the right thing we can include the work everyone wants to see while respecting living artists rights, as well as those deceased but still in copyright. More people seeing art they fall in love with = everyone wins, we think.

And that, is what it’s all about. My philosophy to UX, if I have one, would be best summed up as “Get the groove right, the beats will follow”

* Actually we shifted it by a day – we were one of the launch partners using the new Facebook comments system, and they rescheduled so we all went live on the 1st March together.


22
Jan 11

Week 17 ish

I’m restarting weeknotes! I stopped because I was working on exciting new stuff for our startup, but we were having to be a tiny bit quiet about things and i was bound to say too much. I dislike the ‘stealth’ term for startups, because we are not ninjas or bombers. But sometimes you do have to be a little bit discrete about things, for a short time at least. The last posts were talking about developing our logo, but without revealing the end result it felt a bit pointless. So, a pause.

Now, we’re beginning to emerge blinking into the sunlight. Days are feeling large, the office is buzzing and filling up rapidly. We have done a ton of work in a few different directions (too many at once? Hmm – almost…) but we’re almost there with a one-point-oh. Beta. We have data, we have a website, we have a publishing platform. We have a lot to do.

So not to leave it hanging, here’s a developer  job ad, on githubjobs.

And here’s a logo or two.

Arfinder logos

Artfinder logos

Left we have a reactive logo, that uses colour frequency analysis of your recently viewed content to visualise ‘your’ logo. On the right we have the vanilla.

Enough revealed for this week. More, next.