Surveilled Rather Than Surveyed

Data is great, and it tells you a lot, but it’s cold. A survey where you actually engage is a lot more telling. And like the blood pumping through a hand that you shake -it’s warm.

Netflix revealed, via what some people are painting as a creepy tweet that sounds stalkerish, that they pay close attention to what people watch and how often. Is this really a surprise? Probably not.

Here’s the thing — I watch a lot of programs on Netflix, and I notice the predictive red stars for how much I might like something, and I wonder what the hell they are basing it on. Algorithms are not the be all and end all — corrupted datasets are all too easy to spot. Just imagine if I were to hand out my Netflix password to a varied group of friends (I do not do this), then I might get suggestions that just do not jibe with where my head’s at.

Look at the irrelevant ads that clog the stream on my Facebook wall because I click on friend’s pages to like them, or maybe my browser is clogged with cookies from other people searching around. Just to say data gathered from monitoring a machine or a browser may not be the most accurate way of gauging how someone really feels.

Not to say that these things aren’t useful tools, but they don’t reach far enough — you still need to apply that last filter — human perception — in order to arrive at a conclusion.

Often, what is revealed to you is the what happened, and not the why. Why did they choose to jump that way? With a survey you can better understand the reasoning behind something.

When Windows or Adobe change something without making the chain of logic available that led to that choice it sometimes puzzles me, and it can initially make it hard to wrap my head around the new working method forced upon me. It often seems like they are trying to pull away the table cloth and leave the items still on the table, but this often doesn’t work. The number of people I will see bemoaning the loss of a feature that they used every day to do their job is common, and when companies have responded to this I have generally seen the attitude that they know best.

You will often see a product fail because it emerges into an environment where it solves no problems, and answers no questions. This makes you wonder whether or not any questions were asked, beyond those limited to the life of the designer. An internalized process is useful for a lot of things, but when you have to sell a product, or a service, you have to ask around — and you may have to go back to the drawing board. At least if you have to do that you won’t have jumped the gun and built a rocket ship where you needed a steamboat. Survey rather than surveil, because it will explode problems and open you up to new input from the possible end user that you might never have arrived at normally. Ask, and don’t accidentally kill the golden goose.