Building new things requires measuring them

When starting work on a new initiative, it’s essential to consider how to measure it, if it’s working and if it’s having the desired effect.

That means measuring potential side-effects as well

Considering how it could negatively impact what you’ve already built should also be considered.

In some cases, it’s even more important to consider the potential side effects.

Unintended side effects can not only ruin your existing experiment but what you’ve built so far. They can actively set you back.

For instance, in Supplybunny, we implemented location selection for addresses. The aim was to improve the delivery. But we didn’t consider that it’s a significant addition to the new user funnel and that it could seriously impact our conversion rates.

From then on, we made asking “What harm can it do?” a regular part of our product discussions.

It can make existing things worse

In the worst case, it lowers customer conversions, as in the example above.

Another scenario to consider is where an improvement is positive for new customers but negative for existing customers.

There are also technical considerations. Some features might have a significant impact on performance, maintainability or scalability.

So, ask “What harm can it do”

Asking what harm can it do should make you think more conservatively about implementing new features. I think that is a good thing because most of us are biased towards creating new things a bit too much. Many companies have too many new initiatives that they don’t execute properly.

Considering the potential side effects also makes you put more thought into designing a feature.

For quality-of-life improvements, it’s also a good bell-weather if you should do it at all. If it doesn’t have a huge upside but can potentially harm something, then there’s no reason to do it. If it has a big potential upside, is quick to do, but doesn’t harm anything, then you might as well do it.

If something needs a detailed study into its impacts, there are probably better things to work on.

But it can make other things better too

The converse effect is that you should think about what else can be improved as a result.

The location change mentioned above, for example, allowed us to display more accurate shipping costs across the entire platform because we had the users’ information.

At the start, it’s often difficult to tell the complete impact a change might have. But by regularly asking this question during your design process, you will not only build better and more complete things, but you will understand your business better.