How IMVU learned its way to $10M a year | Hacker News Reader

How IMVU learned its way to $10M a year

venturehacks.com

58 pointssarvesh10 comments

It is a bit long at over an hour but it is great to hear a success story coming after a couple initial failures to connect with who they thought their customer would be. Anyone is allowed to run an A/B test at any time - he says the difference in a product meeting where someone has actual data vs. typical meetings where people are trying to sell their idea feels like breathing oxygen. They ship code 50 times a day - and not just hacking, they have automated regression testing to maintain reliability. All new engineering hires ship code their first day.

agreed. as the costs of collating data goes down, traditional businesses are going to realize that there are a wealth of customer data there that they aren't using. i can't help but think most business will move in this direction.

I can. There is nothing new about A/B testing -- the statistics available to do it have been available for centuries, the techniques for applying it to offline business have been available for decades (direct marketing, principally), and copy/paste Javascript code to do it has been available for years. And still plenty of people have never heard of it.

It is an absolutely amazing tool, one of the most powerful I have available. But its also hard, in a lot of intractable ways: you need to do things twice or twenty times when the first time was "good enough", you need to commit to looking at data and (scary part) making decisions based on it, and you need to learn to accept that reality and your perception of reality are very different things.

"you need to learn to accept that reality and your perception of reality are very different things." -- this can be extremely hard for people to overcome.

How can you change an organization to be more data and results oriented? I do not know. I think that kind of attitude needs to be in the corporate culture, or hammered into people from on high, in order for groups to get past the "highest paid person in the room" decision making process. Does anyone have any advice about this?

Does anyone have any advice about this?

Best Buy apparently did it, and it has been described as a "revolution" spreading from below and a "virus" infecting one working group at a time.

I got my Big Freaking Enterprise Apps shop to take some baby steps into some directions I like by taking the discretion I have as a low-ranking engineer on the totem pole, using it, succeeding, and tying the success to business goals.

I "convert" other engineers by appealing to desire for efficiency and professional growth, and "convert" managers by demonstrating that the quirky "minor little engineering details" we're working on consistently generate savings.

Thanks for the kind words. Always love feedback and suggestions for how to make the presentation better, esp. when giving it to a more technical audience (the MBA's are a little easy ;).

If anyone knows folks that might be interested in hearing (and, more importantly, ripping to shreds) a version of this talk, do let me know.

That was very interesting.

You could spin the part about having new hires ship on the first day off into a talk on hiring.

The part about looking for engineers who've had a project cancelled really resonated with me.

I get a sick feeling in the middle of a project when the management is asking for extra effort, while at the same time making n00b project-management mistakes. It's a feeling of "how dare they waste my effort."

bet Atwood loves that. http://news.ycombinator.com/item?id=481579

Wow. I've always been a little frustrated by Atwood, but this comment just got him removed from my feed reader.

One of the best talks I've listened to in quite awhile. In fact, the last really good talk I listened to was by Steve Blank. No accident there, I'm sure.

I'm in the B2B space and working on applying these lessons to Enterprise SaaS applications. It has it's own challenges and involves a lot more face time with customers but the key lessons - execute faster than everybody else - seems to be spot on so far.

Reply on news.ycombinator.com