Skip to content
Blogs Best-practices

Should I sort out my processes before choosing software?

Should I sort out my processes before choosing software?

Should I sort out my processes before choosing software? We get that question from almost every business owner who is about to buy a new package. The answer is yes, process first. But we mean something different by that than most people think, and that difference determines whether you are happy with your choice a year from now or searching all over again.

The pattern is almost always the same. A company with a team of fifteen people gets stuck in Excel lists, emails and loose agreements. Someone sees a demo of an attractive package, the subscription is signed, and only then does anyone look at how the work fits into it. Within a few weeks the software determines how you work, instead of the other way round. Staff notice that it does not match what they do every day and start working around it. The old Excel file quietly comes back.

A system only does what you put into it. Put a messy process into new software and you have a digital messy process. If you then automate it as well, it just goes wrong faster. And after six months the system gets the blame, while the problem was there before a single licence was bought.

The same applies to business owners who now want to build something themselves with AI or a vibe-coding tool. Whether building it yourself is wise is a separate discussion. But anyone who builds without knowing how the work runs only finds that out afterwards. Then you build it a second time. That is where the real question comes in: what does "process first" actually mean?

In order means mapped, not perfect

Process first does not mean everything has to be perfect before you choose anything. Anyone who waits for that never starts. We know business owners who have been "cleaning up the processes first" for two years and are still working with the same lists. Perfection as a precondition is simply postponement with a respectable name.

Process first means: first write down how the work really runs now, mess included.

You do that with a few simple questions. What steps are there from enquiry to invoice? Who does what? Who owns each step? Where do you wait for each other? Where does it go wrong? And what takes the most time? The first time, that can simply be on a whiteboard in the canteen, or on a Miro board. Allow an afternoon with the people who do the work, not a months-long project with a thick report at the end.

Watch out for one pitfall: write down how the work really goes, not how it should go according to the handbook. The difference is almost always in the exceptions. The customer who phones instead of emailing. The quote that still has to go past the managing director. The rush job squeezed in outside the schedule. Those exceptions are exactly where a new system will get stuck later if you do not know about them. So keep asking the people on the shop floor: what do you do when things run differently from normal? That is where the most interesting answers come from.

With that overview you know what you need. Only then do you choose a system that fits. You walk into a demo with your own list of bottlenecks instead of being tempted by the features the salesperson most likes to show. Ask the supplier to show your three most difficult situations in the package, not their standard pitch. That saves money, but above all it saves resistance from your team, because they recognise their own work in what is chosen.

Three sequences, one that works

In practice we see three ways business owners approach this. The first is system first: choose a package, cram the work into it. That produces the digital messy process we described earlier, with people dropping out. The second is perfection first: spending months mapping out and optimising processes on paper. That sounds sensible, but you are then designing a way of working that has never been tested in practice. Paper is patient, your team is not.

The third sequence is the only one we see working. First understand how you work, then choose what to work with, and after that improve both together. After the choice, process and system move forward together rather than one after the other.

Why improvement only comes during set-up

As soon as you start setting up the process in a system, you naturally see which steps can go or need to change. An approval that three people have to give suddenly turns out to be unnecessary when you set up permissions. A handover between the office team and planning that people always had to wait for can become an automatic notification. You do not get those insights around a meeting table. They only come when you have to decide concretely how something runs in the system, which is why there is no point puzzling it all out in advance.

After that the system does something loose agreements never do: it holds on to the new way of working. An agreement on a whiteboard is forgotten after three weeks. A step that is mandatory in the software simply gets done. That way you improve process and system at the same time, in small rounds of a few weeks, instead of in one big leap that nobody keeps up with.

Actually schedule those rounds too. After go-live, put an hour in the diary every three to four weeks with the people who use it daily. What works, what irritates, what are they still working around? That last one is your most important signal. If someone is keeping their own list alongside the system again, there is a step that has not been set up properly yet. Solve that quickly and the system remains the place where the work happens.

For anyone building their own solution with AI tools, this applies even more. You can cancel a package. Something you have had built on the wrong picture of your work has to be rebuilt. An afternoon at the whiteboard saves you the second build. The overview of steps, owners, waiting points and time-wasters is exactly the brief your build tool needs to make something usable.

So the question of whether you need your processes in order before choosing software is less chicken-and-egg than it seems. Not perfect processes first, and not the system first either. Grab a whiteboard this week, bring in the people who do the work, and write down what steps there are, who owns them and where things get stuck. With that overview you choose more purposefully, your team does not drop out and you build nothing twice. If you would like help with that, Lead Innovation is happy to think along during that first afternoon.

Want to know more about monday.com?

The Lead Innovation team at work