Modern business operations means managing the work, not just the people.
A guide for companies with 10 to 250 employees, without pillar models or buzzwords: where time gets lost day to day, how to put a euro figure on it, and what software can and cannot do about it.
By Jonas Höttler · Updated 22 October 2026 · about 11 minutes to read
A lot gets written about modern business operations: pillar models, customer journeys, emotional intelligence, AI strategies. Most of it isn't wrong. It just doesn't help a team lead much on a Monday morning. She wants to know why the order is stuck with sales again and what that actually costs.
So this guide starts with the work itself. The assumption behind it: in most companies time isn't lost inside individual tasks but between them, in follow-up questions, handovers, duplicate data entry and waiting for an approval. If you want to manage that, you first need a number and then an order of steps.
I'm Jonas Höttler, and with balane GmbH I build automations for mid-sized companies. What follows is the method I work by. The examples are demos and made-up calculations, not client projects.
One sequence instead of three pillars.
Report, measure, cut, automate. The order is the point: automate first and you only take a detour faster.
Report
The people who do the work every day know where it gets stuck. They note a disruption the moment it happens, in a few seconds. Not in the evening from memory, and not in a workshop.
Measure
The notes become a number: which cause comes up how often and how much working time it costs, in hours and in euros. A gut feeling turns into a ranking.
Cut
Before anything is built, whatever nobody needs goes: duplicate entries, approvals without a reason, reports nobody reads. Often a required field or a clear owner is enough. That costs no licence.
Automate
What is left and follows fixed rules is taken over by software, in the systems you already have. Where judgement is needed, a person keeps deciding.
Orange means people judge. Green means software works.
A company isn't a process. It's a network.
Process descriptions usually show one workflow running left to right. In practice the workflows hang together: an enquiry becomes a quote, the quote an order, the order a delivery note and an invoice. Every department works in its own system and still needs the same customer, product and order data.
That's why the expensive spots rarely sit in the middle of a department. They sit at the handovers. Speed up a single workflow and you often just move the jam to the next handover. It pays to look at the whole before you decide where to start.
The map is a model, not an analysis of your company.
What one disruption costs per year.
A ten-minute follow-up question looks harmless. If it happens twenty times a week, that's over 150 hours a year. Nobody does that maths day to day. So the small disruptions stay invisible while the big projects get debated for months.
Try it with a disruption from your own work. The calculation uses the fully loaded cost per hour, meaning salary plus on-costs and a share of overhead, not the gross wage.
Friction calculator
Hours per year
153
Cost per year
€7,667
Hours per year: 153. Cost per year: €7,667.
That's exactly what the friction log is built for: you note every disruption as it happens, and once you have seven observed days and five entries the app projects the monthly figure. Free, no account, the data stays on your device.
Data and systems: what automated data processing does.
Automated data processing means data is captured once and then passed on, checked and posted by software instead of being retyped. A typical example is incoming invoices. The invoice arrives by email, a program reads the amount, supplier and order number, matches them against the purchase order and creates the booking. If something doesn't fit, the case goes to a person.
This rarely needs new software. Usually the systems you already have get connected, such as ERP, accounting, CRM, email and Excel, using tools like n8n or Make.com. AI is one step in this, for example reading a scanned receipt or classifying an enquiry, not the whole system.
What has to be in place
One source per piece of data.
Customer data belongs in one system, not three. As long as the same address is maintained in several places, you mostly automate contradictions.
Fixed rules.
Software repeats what can be described. Where decisions are made differently every time today, someone has to settle the rule first.
A person for the exceptions.
Every workflow needs a place for cases that don't fit, and someone who handles them. Otherwise they sit there until a customer complains.
Customers: where handovers show from the outside.
Customer journey sounds like marketing. For a company with 50 employees it's mostly a series of handovers the customer can feel: how long does a quote take? Is the delivery date right? Does the invoice arrive correct and on time? Do they have to call twice before someone takes ownership?
At these points automation often does more than any campaign. An enquiry that lands with the right contact on its own, a shipping notice with a tracking number, a dunning letter that only goes to customers who really haven't paid: small workflows the customer notices directly.
A/B testing and conversion optimisation have their place, but they need enough visitors to give reliable results. In B2B mid-sized companies with few, large customers, response time and error rate are usually the bigger lever.
Team: who reports, who decides, who runs it.
Automation changes work, and people are right to worry about that. The most honest answer is involvement from the start: the disruptions come from the people who do the work, not from a consultant's slide. Someone who has noted down what slows them down is more likely to see an automation as relief.
Many day-to-day conflicts aren't about personalities but about unclear ownership: who approves? Who maintains the master data? Who calls the customer back? Questions like these can often be settled faster than any team-building exercise, and they have to be settled before a workflow is automated.
One owner per workflow.
They know the workflow, decide on exceptions and run the automation after handover. Without them, every solution is orphaned within a few months.
Goals you can measure again.
Instead of "get faster", try "follow-up questions between sales and production down from 20 to 5 a week". Measure before, and again a few months later at the same point.
Say openly what changes.
Which steps go, which stay and who owns what from now on. Rumours fill every gap management leaves.
Limits: where this approach doesn't help.
The method isn't a recipe for everything. Three cases where automating doesn't help, and one this guide isn't meant for.
When the workflow runs differently every time.
Then rules are missing, and someone has to set them first. Without a standard there's nothing a machine could repeat.
When a step is actually unnecessary.
Better to cut it than to make it faster. An automated detour is still a detour.
When nobody owns it.
An automation without an owner lasts until the first exception. After that nobody knows why it does what it does.
When the question is strategy.
Whether you should open a new line of business or change your business model is not something process measurement answers. This guide starts from the workflows you already have.
The first 90 days, as a suggestion.
This is what a start can look like that a team lead takes on themselves. The weeks are rough guides, not a promise.
Weeks 1 and 2
Report
Pick an area where things noticeably get stuck. Note every disruption as it happens, with a category and estimated minutes. One or two weeks are enough for a first picture.
Week 3
Measure
Evaluate the notes: which cause comes up most often, which costs the most time? Project to a month and a year using the fully loaded hourly cost. What you get is a ranking, not an opinion.
Weeks 4 to 6
Cut
For the most expensive causes, first ask whether the step is needed at all. Add required fields, clarify ownership, get rid of duplicate entry. A lot gets solved here without software.
From week 7
Automate
Pick one workflow that is still left after that, happens often and follows fixed rules. Start small, with an owner who will run it later.
After about 90 days
Measure again
Log again at the same point and compare. Only this number shows whether anything has moved.
Frequently asked questions
What does modern business operations mean for a mid-sized company?
In practice it means managing workflows as deliberately as people: knowing where time is lost, putting numbers on it and improving it in a fixed order. Cut first, then automate, and measure the result again.
What is automated data processing?
Data is captured once and then passed on, checked and posted by software instead of being retyped. Examples are incoming invoices, taking orders over from the online shop, or routing enquiries. Where judgement is needed, a person stays in the workflow.
How do I find out where my team gets stuck?
By having the people who do the work note every disruption as it happens, for one or two weeks. The friction log for iPhone and Android is built for this, free and without an account. You can also assess a single workflow in the self-check.
Does automation cost jobs?
That's the company's decision, not the software's. In the workflows this guide is about, mostly retyping, searching and chasing go away. The time usually flows into work that used to be left undone. What matters is saying openly what changes.
What does it cost to get started?
Reporting and measuring mostly cost attention: the friction log and the self-check are free. What an automation costs after that depends on the workflow and the systems involved. A reliable figure only exists once the workflow has been mapped.
Which workflow costs you the most?
Tell me where things get stuck for you. We look at one workflow together, and I'll tell you honestly whether automating pays off or whether cutting is enough.
You see the short result immediately, without an email address. For the detailed report the check asks for one.