Ask HN: Is it illegal to A/B test pricing?
I've seen the creator of Visual Website Optimizer mention both on his blog and on Quora that a/b testing pricing is potentially illegal.
I understand that it could be illegal if you do it based on user information (gender, location, etc.), but would showing different prices to different users based completely on random selection be illegal?
Note: I'm aware that it could potentially cause a stink if users find out about the different price points, but I'm mostly curious about the legality.
An open Ask HN community consultation from 2011 where @jarin convened hundreds of developers and bootstrappers to share real experiences, hard numbers, and tactics.
Hundreds of active indie hackers contributed case studies, pricing heuristics, and pitfalls to avoid when moonlighting on side projects.
The way you do pricing A/B testing is a bit more complicated.
1. You decide on your range of prices X through Y.
2. Set the 'normal' price to the high-mark at X, then offer users discounts on a scale all the way down to Y.
3. Remember, everyone gets the same discount at the same time. It'll probably give you the most information if you don't discriminate based on any factors.
Now as for legality, you can discriminate based on location---say offering different prices on games to Australians than Americans even though the exchange rate is nearly 1:1.
Or you can discriminate based on gender---say like a club/dating site offers free memberships to women, but not men.
Of course, this is not legal advice per se, and you should see your own legal opinion before acting on anything.
However, legality aside, people will complain if they find out that you're using price discrimination. That's why you're doing it as a "discount". Hardly anyone will complain that someone received a larger discount than them so long as there was an opportunity for them to get it too.
This is an interesting question.
The way I would approach this is by creating a 'sales' page which appears as a 'sale' but actually it is where you are testing your particular price point to see which brings in more conversations.
For instance you can have this "sales" option to load randomly for various users, with different options to test specific price points - and then you give them a different amount of discount (this discount ends up giving the user the price points that you are A/B testing)
Furthermore, to ensure that you avoid any legal lawsuit I would specifically state that that you will honour a lower price price if they find it elsewhere. So if one customer buys it for $50 but another for $40 and they find out, you can just give them the difference.
I'd like to say this isn't legal advice and that I recommend you checking this out from a legal standpoint before implementing but could be the solution to A/B testing pricing points.
I wouldn't. All you need is one person of a protected class to file a lawsuit, where they could potentially find some statistical anomaly in your prices as "evidence" of discrimination.
Why not show different prices on different days (i.e. have sales)... that's definitely legal. Sure people can "game" the system and wait a day for a better price, but those same people would hit refresh / clear cookies / etc. if they found that doing so got them a deal.
Logically you could reduce the A/B interval to hours or minutes and still call it a "sale", but the legal system isn't always logical...
Perhaps the easiest way to avoid suit is to set your minimum price up front, say $10.
Split test the prices up the funnel, but on the actual signup page, only charge $10. Nobody can sue you for price discrimination, and you can track how far down the funnel people make it at various price points.
Once you have the information you want, you can raise your prices across the board.
VisualWebsiteOptimizer.com had a great article on using A/B testing to determine ideal price for a product.
Here you go: http://visualwebsiteoptimizer.com/split-testing-blog/ab-testing-price-testing/