How would you make the tech we already have better? Would you add more buttons to it? A better design, perhaps? A bunch of flashy features is a must, isn’t it?
After all, you want to be able to make a big announcement with all the bells and whistles.
This would all work IF the tech you started to improve actually worked properly.
If not, you’re just wasting your time because no amount of features can make a non-functioning base work. But even though this is both logical and obvious, this happens far too often. Just think about all those failed autocorrect features or supposed “smart” devices that aren’t that smart at all.
The reason all this happens is that the data that’s running under all of the flashy stuff isn’t good, and when that doesn’t work, nothing else works.
Why Features Usually Win the Roadmap
Do you know what the most valuable resource is today?
Gold? Oil? Perhaps diamonds? None of these. It’s data. And developers are very aware of it.
Nobody is making bad choices on purpose. It’s not like there’s a big conspiracy or that there are some companies creating something that doesn’t work on purpose, and then they try piling features to hide it (cough Todd Howard, cough Sean Murray cough).
This happens because companies focus more on pushing new products out instead of fixing/updating what’s already out there.
'Shiny' Prototypes
If you want to create excitement or do a pitch, it’s super easy to make a barely functioning (or non-functioning) prototype, throw that into a meeting, and say, “Hey, look, I made this.” And from the company’s perspective, this is a pretty big deal.This has to do with exciting new features usually getting crazy amounts of attention because of hype and trendiness.
This is progress everyone can clearly see, and investors absolutely love it. It fits into schedules and gives everybody involved a clear win they can celebrate.
If you work on data, though, that’s invisible.
Most Data Is Assumed to Be 'Good Enough'
You plug in a data feed, it works, you move on.
You don’t stop to ask if the information is actually useful and correct, especially if it’s coming from a partner company or a public source.
This problem behaves like a slow leak, and you don’t exactly notice it until it’s too late. If you end up waiting until the data starts showing obvious issues, you’re (likely) already in trouble.
Real Life Is Messier
The world is messy and unpredictable. And since some software is built for this messy and unpredictable world, you have to figure out how to deal with things changing all the time and keeping up with all the chaos in a swift and accurate way.
If you have unusable, incorrect information flowing into the app, then it doesn’t matter how many cool features you build on top because nothing will work.
Think about the scale of life and the speed at which it all happens. Don’t you think reliable data is a must-have?
How Weak Data Causes Even the Strongest Systems to Fail
Let’s say you just built some truly amazing features on a sort of shaky foundation.
What happens next? At first, you won’t notice it. Then, there will be a hiccup here and there. A number will feel a bit off, a recommendation will seem strange… Stuff like that.
It’s not ideal, but it’s not something to sound the alarm for.
The problem is that you’ve built a system that’s connected, so no tiny issue can stay tiny forever. An automated process will pick it up, an algorithm will multiply it, and then it gets passed along to whatever’s next. What used to be a single piece of data will ripple out and cause the system to make a much bigger mistake at a different place. The biggest issue here is that you won’t know that something’s wrong until you notice a failure way downstream.
It’s even worse if you depend on information from the outside world.
You can’t simply guess or smooth over some things.
Here’s a great example of why:
Think about a community health program.
Their schedules have outdoor appointments that are based on forecasts, which makes total sense. Or take a farming tool that calculates water needs; it can’t do it without a reliable source for weather data, right?
It’s the same with socio-economic conditions – if your base information is outdated or wrong, then you can expect every single output (decision, prediction, analysis, etc) to be wrong as well.
Conclusion
Companies seem to be competing about who can throw out the new shiny feature first, instead of fixing up what’s already here.
If we have something that works, then we might think about focusing on that, instead of creating something new, just because it’s (sometimes) profitable.
At the end of the day, wouldn’t you rather have a weather app that shows correct/accurate real-time weather conditions instead of one that’s super shiny and pretty to look at, but provides you with bad/inaccurate info?
Seems like a no-brainer.
