Tuesday, February 22, 2011

Eat your own dogfood (pairs with Merlot)

In the last hour, the tweet from @danmartell  reminded me of the most important lesson I've ever learned -  and learned the hard way. Ironically, I also heard something similar from @jason Calacanis last week at the Startup Grind meetup.

Dan's words are below, and Jason's words were "if you and your team don't use [your products] everyday, something is really wrong", and he went on to talk about how he made excuses for Mahalo 1.0 on why he didn't use it, but the signs were all there that there was something not right with Mahalo 1.0, and the key was that he wasn't using it everyday. It was a very honest and enlightening discussion, and brought back memories. You can see Dan's similar comments in this picture:

If you are not using your product, you're dead-company-walking.

Although it's often attributed to startups, the honest truth is, it happens to everyone. My experience was during my time at Apple - yes, that Apple - even Apple built stuff people didn't want, once upon a time.  I'm not going to bore you with my personal backstory, but I was at Apple as a 20 year old, first working on MacApp (not the new MacApp, but a C++ framework for applications), even getting to speak about it on stage at WWDC. I then transfered over to be a tech-lead on a OpenDoc subproject called A.L.O.E. (Apple Library for Object Embedding).

Jokes about naming aside (ALOE? Remember, this is the company that brought you iPad), these were both killer jobs, where I worked with the most amazing people directly and indirectly, got my ass kicked when I wrote crappy code, and learned to be a "real" developer.

MacApp, as you can see from the Wikipedia article, had a long and successful history that I'm proud to be part of; OpenDoc - which I'm also proud to be part of, was also full amazing people, great idea, well funded, and even the same floor of the same building as the MacApp team. OpenDoc was, idea-wise, well ahead of it's time, and can be seen, conceptually, be found in most current technology. Just last year, O'Reilly posted an article about how the ideas of OpenDoc would improve the iPad experience. It was an amazing project, full of immensely talented people. But it's life was short.

I should be clear, I was not part of the core OpenDoc team like I was for MacApp3, so my point of view may be a little different than others; ALOE was a subproject. Despite what the people say about OpenDoc's failures, I think the biggest issue was we didn't use OpenDoc, or any of the products our developers or customers created.

When someone says your product is slow, that's a pretty good sign you are not using it everyday.

Sure, a few people used Cyberdog, but in general, the team didn't use OpenDoc parts such as WAV to keep meeting notes, we used Microsoft Word. We didn't use Numbers and Charts, we used Excel. None of our Apple internal tools (such as Radar, our bug tracking tool, built in...MacApp) was adopting OpenDoc.

We didn't build applications or parts we used every day, but we build a lot of demos. We should have used, every day, the things Thomas Wiesbach (and Steve Smith's team, if I remember correctly) developed; we should have built things we used (ala 37signals). Instead, we made demos and demo code. I was no better, I should have been building mashups showing how to use the tools in way I would use them, everyday, but I did not. In retrospect, it felt more like a science project to me than a product development process.

In retrospect, and worst of all, we didn't support our developers by using their products, by giving them usability and product feedback. We interacted with them in marketing and  technically (and our support team was great for technical bits), but not religiously, not as a band of people all trying to change the world. Culturally, we got that wrong. That should have been part of our manifesto, like it was on the MacApp team, where we constantly created applications, got into the developer community every day, and "ate our dogfood."

Claris did make some noise, nothing shipped, unfortunatley. If you read that linked closely, you'll see that there is Apple's wholly-owned application arm, Claris, saying "OpenDoc doesn't fit our business model". Um, time to listen to the customers! That said, there is argument that Claris actually didn't like the product technically, and Apple tried to force it down their throat. I wasn't part of that directly, although I did support Claris in some Object Embeding. In particular, Tantek (a very smart guy, by the way) moved mountains for Object Embedding and Claris. In the end, I have no idea if this is accurate or not.

When Steve Jobs came back, he specifically said "we put a bullet in OpenDoc's head".  It cost me my [at the time] dream job, as nearly the entire team was let go. The good news is, I learned a lesson very early on that I've tried to apply to every startup, every project, every idea since then. Sometimes it's hard. Sometimes you are building something for someone else (like we did with the Health Buddy), so you need to find a proxy, and do endless UX review and user feedback.

But you can't over look the simple fact - if you don't use it, why should anyone else?