Ask HN: How do you use beta testers?

https://news.ycombinator.com/item?id=1990973

by TheBurningOr • 16 years ago

12 1 16 years ago

news.ycombinator.com

Does anyone in the HN community use the old paradigm of the beta tester? That is, do you have a select group of tech-savvy, early-adopter, high-feedback users that you use to test new features or new products? OR do you just throw a MVP out there and A/B test every iteration on all your users.

The discussion earlier today about 'toxic clients' got me thinking that the reverse could be true too and that there could be a class of, we'll call them 'gem clients,' who not only use your product with minimum fuss, telling their friends how great it is, but who also provide valuable feedback and suggestions? It seems to me the idea of a beta-tester still has value in-so-far as you can use it to get these high quality users early on, lowering the feature-set and polish necessary for a product to be minimally viable and providing real-world feedback at an earlier stage in your product development.

So how do you get those customers to adopt your platform? Where do they come from? Are they as valuable as my musings think?

patio11• 16 years ago

I generally do a preview release for techie friends, but I've never had a formal beta for BCC or AR. If the software is stable enough to provide value, then I release it and start charging money. You get much better feedback when someone is invested in your software and wants to use it to get their work done, versus someone who is not invested in your software and for whom using it is uncompensated work.

(My conscience is totally clear about doing this: I give refunds for any reason as a matter of policy, and if people didn't sell buggy software then there would be no software on the market. In return for higher than average suffering compared to later users, they get me at their beck and call for the first few weeks. In the very unlikely event that AR failed to phone the customers' customers, I am more than capable of calling Lucinda Smith to tell her that her appointment with Mindy's Hair Salon is on Tuesday at 3 PM.)

BCC 1.0 was terrible software. Somebody bought it anyhow. (He got a refund.) 1.02 was heavily influenced by his comments, and sucked a little less.

AR 1.0 is not terrible. It is bare-bones, and it certainly isn't right for all customers, but it does what it says on the tin. It will probably improve fairly rapidly with customer feedback.

I get users for software four weeks old the same way I get users for software four years old: organic search. (There is a reason I got AR's website up and some links pointed at it six months before I actually had software to sell: I now get a couple hundred searches a month "for free", which gives me a small trickle of brave souls to try it out on.)