Business process automation: where to start so it makes sense and does not cost more than it saves

A manufacturing company's director called us saying he wanted to automate everything. He had a list of 23 processes he wanted to digitise. We ended up with three. Those three will save him 35 hours a month and pay back within six months. The remaining 20 would cost more than they would ever save. This article is about how to make that choice correctly from the start.

Why most automation projects fail or disappoint

Automation fails for two reasons. Either the company automates chaos, or it automates a process that does not repeat often enough. Automated chaos is still chaos, just faster. And automating a process that runs twice a month and differently each time will never earn back its development cost in five years.

Proper automation starts differently: not from the technology, but from analysis. Where do your people spend their time? What repeats? What is precisely defined and what changes a little every time?

How to identify processes worth automating

The repetition and definability rule

A good automation candidate is a process that happens at least once a week, has clearly defined steps and always produces the same type of output. Incoming emails with invoices, creating contacts in CRM after an enquiry arrives, forwarding orders from the e-shop to the warehouse, generating weekly reports.

A bad candidate is anything where the outcome depends on the judgement of an experienced person, where the context changes case by case or where an error could have serious consequences without human oversight.

The hourly log method

The simplest way to identify processes worth automating is to ask every employee to track for one week what they are doing and how long it takes. The results almost always surface three to five processes that are both the most time-consuming and the most repetitive. Those go first.

Where to start: three process types with the best return

From our experience across dozens of projects, three areas deliver the fastest return:

• Email and communication: sorting incoming messages, automatic replies, escalating urgent enquiries, creating helpdesk tickets
• CRM and leads: automatically creating contacts from forms, assigning enquiries to sales reps, follow-up reminders
• Documents and reports: invoice processing, generating statements, archiving supporting documents

These processes are well-defined, highly repetitive and their automation does not require changing how the company works. It simply removes the manual transfer of data from one place to another.

What to do with the results of automation

Automation by itself is not an outcome. The outcome is what your people do with the time that automation frees up. If a sales rep stops manually copying contacts into CRM and spends that time with customers instead, automation shows up in revenue. If they spend it browsing the internet, nothing has improved.

That is why every project we deliver ends not just with handing over the solution, but with a recommendation on how to use the freed-up capacity. Automation is a means, not an end.

Want to know which processes in your company make sense to automate first? Book a free company audit. We will go through your processes and prepare specific recommendations with a return-on-investment estimate.