Famously, this is what Justin.tv did. They split the company into three parts:

1) Justin.tv 2) Twitch 3) SocialCam

I suspect if they had to choose one, they would have done SocialCam, and they would have missed the billion dollar opportunity that was Twitch.

Like most things, it depends.

We did both for a relatively short period of time (about 1 year) when we were a profitable company. So... very special circumstance. Btw - its kinda cool that people even know the backstory to Twitch.

HackerNews proving once again that you can get incredible insight in real-time from names big and small :)

Credit to the team having both you and Justin as great story tellers.

I remember listening to you guys talk about your experience starting the company on a podcast (I think it was Startup from Gimlet media?). It was an awesome episode (I think two actually) and was very inspirational to listen to. I love the part when you guys received the money during one of the founders' wedding and seeing the number on the bank account.

(my personal experience) 80% of the time, do both is the wrong choice. 20% of the time, it's the right thing. So say no 80% of the time and say yes less than 20% of the time.

Another version of "do both" I have seen and been guilty of is "let's just keep supporting X in case it will become relevant in the future".

Actually sitting now and deleting part of the code that should have died a long time ago.

Oftentimes you can frame it as "try both" successfully. If you have a big decision, think of the smallest, quickest way you could test both of the options and get some visibility on the results. Then do that, pick the one that's working better, and stop doing the alternative. You can at least be open to unexpected successes with this strategy, and protected against unexpected failures.

This also has the advantage of forcing you to think of the smallest, quickest way to test your hypotheses, which has a fair amount of value itself.

Yeah definitely agree here. This is probably also the best way to actively to avoid "doing both" in the long term. Twitch does seem to be a great example here (as mentioned above) and Emmett also talks a bit about it in his user testing class from startup school (starting @ 2:40ish).

https://youtu.be/qAws7eXItMk

I have mixed feelings on that advice. Sometimes, keeping a pair of old clients that have a closer relationship with your company provides priceless feedback on the product.

The advice is, in general terms, correct. It is not an absolute, though.

I really wish more information was written about why "do both" was so bad.

Why were the high touch customers such a distraction?

Because they're high touch? White glove versus self service. Self service Will always scale with technology, white glove services only scale with more white gloves.

Depends on margins. White glove works fine as long as margins are high enough. Many of the largest clients are not going to do self service.

"if it doesn't hurt, it isn't a strategic decision"

I suppose it may vary, but what do you think is the right way to decide on which of the non-do-both options is best? The one with more automation? The one that maximizes runway?

Of course it depends, but this is where having a mission statement is really helpful. I've been in this position before and what seemed like a hard decision at first was super easy once I reread the mission statement. "Oh wow...option B clearly fits into the mission while option A does not."

I would encourage every company, large or small, to work on a real mission statement that can be committed to for the long haul.

As I have gotten older, this is a good way to judge potential employers, especially if they cannot articulate the mission.