You Can't Automate Everything at Once — Here's What to Fix First

The problem isn't finding things to automate

Most small business owners we talk to don't have a shortage of automation ideas. They have a list of ten, fifteen, sometimes twenty things that feel broken — invoicing, lead follow-up, onboarding, reporting, inventory updates. The problem is they don't know which one to tackle first, so they either freeze and do nothing, or they start with whatever's most annoying that week.

Annoying isn't the same as expensive. The process that irritates you the most is often not the one costing you the most money. You need a way to rank candidates objectively, or you'll keep automating the squeaky wheel instead of the leaking pipe.

A scoring framework that takes 20 minutes

For every process on your list, score it on three things:

1. Hours per month — how much human time does this eat up, across everyone involved?
2. Error cost — what does it cost you when this goes wrong (refunds, rework, lost deals, compliance fines)?
3. Build effort — how hard is this to automate, roughly, on a 1-5 scale (1 = a single Zapier flow, 5 = custom API work across multiple systems)?

Then calculate a simple priority score:

Priority Score = (Hours per month x hourly cost) + Error cost, divided by Build effort

Higher score, higher priority. This isn't meant to be scientific — it's meant to stop you from automating based on gut feeling alone.

A worked example

Here's a composite scenario based on patterns we see across clients — a 12-person marketing agency trying to decide what to automate first.

Process	Hours/month	Labor cost (at $35/hr)	Error cost/month	Build effort (1-5)	Priority Score
Manual invoice creation	14	$490	$200 (late billing)	2	$345
Lead follow-up emails	20	$700	$1,200 (lost deals)	2	$950
Client reporting (dashboards)	10	$350	$0	3	$117
Onboarding paperwork	8	$280	$150	2	$215
Inventory restock alerts	6	$210	$600 (stockouts)	4	$203

On pure hours saved, lead follow-up and invoicing both look attractive. But once you add error cost — in this case, slow lead follow-up quietly killing deals — lead follow-up jumps to the top by a wide margin. Reporting looks tempting because it's visible and annoying, but the math says it should actually be fourth in line.

This is the part people skip. They automate what's visible, not what's expensive.

Why build effort matters more than people think

A process that saves 10 hours a month but takes three weeks and $4,000 to automate properly has a very different payback period than one that saves 8 hours a month but can be built in two days. If you only look at hours saved, you'll consistently pick the wrong first project and get discouraged when the ROI takes six months to show up instead of six weeks.

The rule of thumb we use: your first automation project should pay for itself in under 90 days. If it doesn't, it's not a bad idea — it's just not a good first idea. Save it for round two, once you've got a working system and some internal buy-in from a quick win.

What this looks like in practice

Using the table above, the agency's actual rollout order was:

1. Lead follow-up automation (score: 950) — built in 9 days, paid back in under 6 weeks
2. Invoice automation (score: 345) — built in 4 days, paid back in about 7 weeks
3. Inventory alerts (score: 203) — built in 3 weeks, paid back in about 10 weeks
4. Onboarding paperwork (score: 215, but deprioritized due to a dependency on a CRM migration)
5. Reporting dashboards (score: 117) — last, despite being the most requested internally

Notice that the thing everyone complained about most — the manual dashboards — ended up last on the list. That's usually how it goes. Frustration and financial impact are not the same metric, and conflating them is the single biggest reason automation budgets get spent on the wrong things.

The takeaway

Before you automate anything, write down every candidate process, score it on hours, error cost, and build effort, and rank them. It takes less time than most people spend arguing about it in a meeting, and it turns a gut-feel decision into a math problem — which is exactly where automation decisions should live.

If you want a second set of eyes on your list, or help scoring the build-effort column realistically (this is the part people consistently get wrong), that's a conversation worth having before you spend a dollar on tooling.