Why Process Mining Got Mistaken for an Enterprise Tool

When we talk to experienced process mining practitioners about our target market for Flow Myna, we almost always get the same pushback: smaller businesses won't have the volume of transactions to make it useful.

It always makes me stop and think. Because I've worked on use cases with a handful of event types and modest transaction volumes where process mining was genuinely valuable. But we were only doing that because a few years into our process mining journey we'd struck an enterprise licensing deal, and we had a team of skilled process mining engineers and analysts who could tackle any use case in-house. No third-party process mining implementation partners needed. No reliance on off-the-shelf system connectors for data collection and data transformation.

Most businesses never push into those off-piste use cases. They are stuck with the standard connectors that serve high-volume, high-value processes only. The ones the process mining business case was built on. The ones that justified high license costs.

Built for the Big Deal

The process mining market has historically been dominated by a few big players with expensive license costs and off-the-shelf connectors for big enterprise software (think SAP and Service Now). That's gradually shifting, but it means that process mining has mostly been deployed within large enterprises. This also means that those are the use cases that get written up, presented at conferences, and headlined. Big ROI. Working capital freed up. Procurement cycle times slashed.

These players aren't publishing case studies about how process mining improved visibility into a coordination-heavy process, avoided the costs of hiring improvement consultants and empowered employees to act with confidence. These aren't compelling enough stories when the typical cost of a process-mining license is a six or seven figure sum.

So the mental model of process mining has been shaped by what's visible. Big enterprise. High transaction volume. Only used when there are millions of dollars on the table.

In my previous role, the economics of this model became clear when HR came to the table at the start of our process mining journey. HR processes, especially when you're managing employees on a global scale, are legitimate, high-value processes. But the transaction volume is structurally capped by headcount. At any given time, you're not going to get close.

And the ROI for onboarding HR? The benefits were implicit, but difficult to quantify in the way a CFO expects when investing that much in a license. Working capital initiatives have a clear dollar return on every day of improvement. HR processes don't work like that. But if you're losing talent because your processes are scrappy and difficult to coordinate, your business is taking a hit, it's just difficult to hang a price tag on it.

When the license cost is high, you need a quantifiable ROI to justify it. When it isn't, you can treat process mining the same way you treat any other analytics infrastructure. Something you use because it is the best tool to help you understand how your business runs, and it will enable subsequent projects to be successful.

We spent over a year negotiating the commercials for HR because the license model didn't know what to do with it. In the end, we bundled it in with an on-prem to cloud migration deal. These smaller processes came along for the ride during a system upgrade.

That's the world process mining has grown up in. Volume and scale have become the mental model for who it is for. Not because the technology requires it, but because the license model and available system connectors did.

Another Use Case That Defied the Playbook

I once worked with a team to visualise their global shipments process. A handful of events, very few variations, and maybe a hundred active shipments at any one time. Nothing like the classic order to cash or purchase to pay use cases. And yet one of the most successful implementations I've been involved with.

Tracking shipments as a visual graph showing what has happened and what state they're currently in, does something a table of rows doesn't. There's a reason every tracking app, every order status page, every CRM pipeline uses a linear flow to represent a single object in motion. Your brain interprets sequences naturally, because that's how it models causality. A chain of events maps directly onto how you already think. A status field gives you a snapshot with no context. A table makes you reconstruct the sequence yourself. A flow just shows you.

Process mining extends this to the aggregate. Not just one shipment, but all of them overlaid. Median and average times sitting right on the process graph. Where flows deviate from the normal path and how often. You're not reading numbers and mentally mapping them onto a process. The numbers are on the process graph. And your brain has an easier time grasping what's going on than it ever would from a table, a chart, a status column, or a heat map because the shape of the information matches how you think.

And maybe because the volume was low, the process map stayed legible. One of the real barriers to process mining adoption that is rarely talked about is that the discovered process graph can become such a hairball of complexity that nobody knows what to do with it. When the process graph is small enough to understand at a glance, people use it.

“But Don't Smaller Businesses Just Know Their Operations?”

The flip side of the volume argument is that smaller businesses are close enough to their operations to just know what's going on. With AI making it easier than ever to spin up basic charts, is process mining solving a problem they don't actually have?

For some businesses, at some stages, this may be true. But it is likely that the moment the business starts growing, the gap between what management thinks is happening and what is actually happening opens up faster than anyone expects. The founder who knew every customer order by name is now running a team.

And the charts, however easy they are to build, still only tell you what. Not how or why. This distinction matters. Process mining and traditional analytics are solving fundamentally different problems. Getting from raw operational data to time-aware flows you can both visualise and query isn't a prompt away, it still requires software built specifically to structure data in that way.

Final Thought

Volume and scale have become the mental model for when process mining is useful, in part, because of how it was historically priced. Plenty of processes were priced out because the value of visibility was difficult to quantify against the license cost. And plenty of transformation programs never got the benefit of process mining because process mining was operating as its own program with its own business case.

The sweet spot is when process mining can be treated like any other tool in your analytics infrastructure. Not something you need to justify with a business case every time, but something you reach for when it's the right tool for the job. If your numbers are the result of a process, process mining is probably that tool.

That's why we're building Flow Myna. We believe that seeing aggregate metrics overlaid on a map of your day-to-day business flows unlocks a business understanding that is more difficult to get from numbers in a table or chart. Using this as a selling point hasn't been possible due to the high licensing costs and standard connector model which has shifted the mental model to high-volume, high-value processes only.

www.flowmyna.com