It’s the start of a new year, which means: Prediction time for a lot of people and organizations. I have read many, many technology predictions over the last couple of weeks: Some good ones, but unfortunately also many (very) bad ones. I say unfortunately because most of them made mistakes that are easy to spot and avoid.

Here are three things to look out for when making predictions around potential technology futures.

Desirability is not inevitability

We all have ideas how technology should work and features we’d like to see — even if the technology does not yet exist. We also have our favorite technologies that we use daily, are financially invested in, or just find neat: Some of us are really into mobile phones, or machine learning, or VR. That’s OK but projecting your desires onto your favorite technology does not make a prediction.

Beware of doing this:

“I want a [Technology X] to do [Feature Y], so [Technology X generation +1] will allow that.”

Examples I’ve seen:

This is not a prediction, it’s a form of wishful thinking.

Instead, explore why you want the respective technology to evolve in a certain way. Find the core of your desirability and which audiences this might also apply to. Examine if this audience is big / relevant enough as a business or societal driver for the respective technology to evolve in this direction.

Also, always make sure the technology can progress in this direction in the first place, otherwise you end up with a golden hammer.

Familiarity is not a trend

Technology enthusiasts know their way around spec sheets. They know what the processor model numbers mean, generation codenames, product family denominators — the latest and greatest for each technology category. That’s great, but not always a good basis for predictions.

Don’t do this:

[Technology X] is currently the most powerful option in [Category Y], so a new [Category Z] must also use it.

Examples I’ve seen:

This is sometimes also referred to as “Top Trumps Prediction”, where the author just takes whatever highest stats exist for each component and compiles the winners into “a new innovative technology” or product.

To turn your knowledge into a prediction, start with the technology you are making the prediction about. What is its current status quo? What components does it rely on and what are meaningful constraints and limitations (for example size, power, heat)? What are issues, gaps, and challenges — what is the technology missing? And then apply that information to potential innovation paths for individual components within the limitations.

Disruption is not linear

There are multiple types of innovation, usually distinguished by the impact they have on an industry or an audience. Incremental innovation is usually a direct improvement of something existing, radical innovation is usually a major change of something already existing and the often-used disruptive innovation is usually a functional, significantly better substitute, based on something completely new. The problem is that everyone claims to be disruptive these days.

Stay away from:

I think [Technology X] will make a big jump in capability, so it must be a disruptive innovation.

But just because something is new it’s not necessarily a disruption.

For a prediction, try to dive deeper: What really changes? And for whom? Can they transfer existing skills and knowledge? Is the technology relatable or comparable in any way or is it really something completely new and alien to the industry or audience? What are the risks associated with being too familiar or not familiar at all? What will adoption look like? What happens if it replaces the old — during and after displacement?

And finally, some general things

Here are some general guidelines that I find helpful for predicting technology futures.

I hope this encourages you to write your own predictions & being a little more aware of some of the pitfalls. Have fun exploring potential futures and let me know if this was helpful.