The uncanny power of small utilities

We tend to gravitate towards big projects, but small tools can have a large impact in aggregate

April 24, 2021
881 words (5 min read)
RPA Program
Process Design

Small utilities

Utility is an economics/philosophy term that describes roughly how much you value something. Certain things tend to be assigned large utility (e.g. cars), while others are more modest (e.g. new shoes). Normally, utility correlates with willingness to pay, in that we should be willing to spend more on things we value more.

This post is about the accumulating effect of creating tools for small improvements in your (professional) life, which is easy to underestimate. Small utilities aren’t flashy, and often times it’s even hard to get them funded, but if you combine enough incremental improvements, they can add up to something significant.

What’s the utility of an automation?

The elephant in the room is, of course, saving time. If you automate something, you won’t have to spend time doing it, which is good because you can then focus your attention on other tasks (become more efficient) or reap the benefit by reducing work hours (reducing staff overtime or staff over time).Horrible pun? Feel free to unread

But beyond mere time savings, automation can have utility in several other ways:

  • Reduce errors
  • Consistency
  • Get rid of unloved tasks
  • Reaction speed (customer service)
  • Do tasks more often (e.g. reporting)

These other effects are often less tangible than the direct time savings, but no less real or important. One of my favorite automations, for example, handles synchronizing time bookings between two different systems. Copying things from one system to another is one of the most will-of-life-draining activities I can imagine, coming very close to digging holes and then filling them back in. This would be an example for both saving time and getting rid of unloved tasks.Of course, it is debatable whether we should have two systems where we need to enter our times in the first place, but that’s a different discussion

Of small tools

To get to the real meat of this post, many of the automations I use most often confer only very little utility — for example, translating an e-mail from German to English or vice-versa when I want to forward it from our customer/product team to the other side.

Each individual use of such a small task automation only saves half a minute here or there, but these seconds and minutes add up over time. And what’s especially interesting is that small automations like this are more flexible and can be applied in more situations. It’s a bit like a Swiss army knife: not very good at any particular job, but you can carry it with you everywhere.

Compounding utility

Due to their high degree flexibility, small tools that help you with your daily tasks can often build on each other, multiplying any efficiency gains into a compounding effect. If I can do 20% more work because of automation A, and 20% more because of automation B, my output increases by 44% — not 40% (1.2 * 1.2 = 1.44; real numbers will be smaller, but you get the point).

Attended task automation

Most such automations, I’ve found, are of the attended type, and often even only applicable to a small circle of citizen developers and their teams. They turn you into a cyborg of sorts by offloading parts of your work to the robot, which only works if you can interact with it seamlessly.

Because the benefits of automating small tasks are difficult to predict and measure, as well as quantitative and accumulative, it tends to be much harder to gain approval for them. Hopefully, now you know better than to dismiss their utility out of hand, however.


Because they do just one tiny thing, development for small tools is often quite fast. So even though savings may be small for each task automation, if you can keep overhead small, too, you can still get a pretty good return on investment. The risk of cost overruns is also smaller because it’s more difficult to forget half of your process if it only involves 5 clicks and an API call.

Keeping overhead small is another reason why I’m a fan of decentralized citizen development. Citizen devs can cut some corners to quickly come up with something that works for them and their colleagues, but wouldn’t be robust enough to be rolled out to the whole company.


There are, of course, more challenges beyond merely getting funding:

  • It can be difficult to predict which tasks can be usefully automated, especially when machine learning is involved
  • You need a large number of users and/or do the task quite often to justify the development effort
  • Opportunities are easily overlooked and can, therefore, be hard to find
  • Too much governance can ruin your ROI


All things considered, however, creating small tools to make your and your colleagues’ lives a little easier is a very worthwhile endeavor. I hope this little essay got you thinking about some ideas how you might elevate your RPA program by a thousand tiny butterfly wings. Until next time.

© 2021, Stefan Reutter