Here's a question worth asking: how many hours did your team spend last week moving data from one system to another?
Not building things. Not serving clients. Not winning new business. Just copying, pasting, chasing, and updating fields that could — and should — update themselves.
For most businesses with 5–50 staff, the honest answer is somewhere between 10 and 40 hours per week. Often more.
Why it happens
Manual processes are sticky because they start as sensible shortcuts.
Someone needs to get a quote to a client fast, so they copy the data from the CRM into a Word doc manually. It works. The quote goes out. No one builds the automation because the fire is out.
Six months later, three people are doing that same copy-paste job every day. It's now a "process." It has a person responsible for it. Questioning it feels like questioning the person.
The second reason: most small business owners don't know what's automatable. If you've never had a technical person map your workflows, you don't know that the 45-minute monthly job of pulling sales data into a spreadsheet could run automatically in 30 seconds with a £0-per-task n8n workflow.
The real cost
Manual processes have four costs that rarely show up in a budget line:
1. Time cost. An hour of manual data entry is an hour not spent on billable work, client relationships, or growth. At £40/hour fully-loaded, 20 hours/week of preventable admin is £800/week — £40,000/year.
2. Error cost. Humans make mistakes. A mis-keyed figure in a report, a missed renewal date, an invoice sent to the wrong address. Each one costs time to fix and sometimes costs clients.
3. Bottleneck cost. Manual processes create bottlenecks. If one person knows "how to do the month-end," you can't scale, can't take holidays without things breaking, and can't hand off cleanly.
4. Talent cost. Good people don't stay if their job is to be a human spreadsheet. Turnover is expensive and often traceable back to roles that should have been automated.
The fix: map before you build
The mistake most businesses make with automation is jumping straight to tools. They sign up for Zapier, build a few simple connections, hit the limits of what Zapier can do, and conclude "automation doesn't work for us."
The right approach is to map first:
- List every recurring task your team does that involves moving, formatting, or sending data
- Rate each by frequency and time-cost
- Identify which ones are rule-based (same logic every time) vs judgment-based (requires a human decision)
- Prioritise the high-frequency, rule-based ones first
Rule-based tasks automate well. Judgment-based tasks don't — but often the data preparation around them can still be automated, so a human makes the decision faster with better information.
What good automation looks like
Properly built automation is invisible. It runs in the background. Data appears in the right place. Reports generate themselves. Notifications go to the right people at the right time.
You know automation is working when a team member asks "who sent that report?" and the answer is "it just runs."
The tools we use — primarily n8n — are self-hosted, which means you pay for compute, not per task. At scale, this is 80–90% cheaper than Zapier or Make. And because it runs on your infrastructure, your data doesn't go through third-party servers.
Where to start
If you're wondering where to start, pick the most complained-about task in your business. The one people sigh when they mention. The one that always falls through the cracks because everyone assumes someone else is doing it.
That's your first automation. Map it, build it, watch it run.
If you want someone to map your processes and build the automations for you, that's what Deplexify's Automation service is built for.