The Technology Placebo Test
When I was rethinking the other day about how much more I’ve been using static HTML instead of big-blown-up projects lately, it dawned on me that a lot of what we do in our daily life, the projects and things we work on barely have an impact on people nor make their lifes easier. Of course projects like the lean business canvas already try to cover all aspects to figure out, whether your project is worth the time you’d be spending on it. But especially for technology their canvas is only partly suited to figure this out. So I’d like to propose to do mind experiment before you start to develop your technology, I call it the “technology placebo test”.
You probably know about the placebo test in medicine, correctly referred to as the Placebo-controlled study. What you do is you always have a control group of patience which get a similar looking and tasting drug that has no effect, the placebo, within your group of patience. When you want to bring a drug to market it has to significantly stand out in the results in comparison to your placebo-control group. Now, we do not really have any placebo in technology - or do we?
Maybe a similar effect exists with technology as it is exists on medicine. After all it is only the believe that we are so advanced and the drug has to work that makes the placebo work for so many people. What if there is a similar effect on technology, our believe that we create superior technology overlooking major flaws and just be being told it solves our problem we believe it actually does. Let’s take a recommendation system for example: If you have an advanced user, who believes the technology underlying is gonna solve his problem, he’ll probably see the results that way. Even if those would just be random.
So if you want to figure out whether your technology actually solves the problem, try to not only gather feedback from an A/B-Tested group of users, where one groups didn’t get any advancement at all. Like in the case recommendation case, push it out to users and some receive a random result (or the ones from your old algorithm) while others receive the new one. And you ask everybody to rate the results. Now you have a real case, on whether what you are doing is actually better or just because people believe so.
Btw. when I was doing support in a small tech shop when I was a teenager, we had saying towards our customers: if it is possible without computers do it without. Like the guy who needed a copy of a paper and was thinking of buying a full-blown-scan-fax-printer. We told him to get into the next copy-shop and get a copy. Easier, quicker, better. So always also look into the offline world whether what you are building is actually better in the proposed case; does it really save people time? Does it make their life easier? Or just more complex? Think about how users would rate it, if they were just told it is quicker and easier though it ain’t. Make the technology placebo test.