· 2 min read · By Taha Rahmani
What Should You Automate First? A Practical Framework for SMBs
Not every process is worth automating. This practical framework helps you rank workflows and compare a proposed build with a measured operating baseline.
The first automation project matters more than any other. Get it right and you build momentum, trust, and a budget for the next one. Get it wrong and “automation” becomes a dirty word in your company for two years.
The good news: choosing well is mostly arithmetic.
Westflow's four-factor scoring framework
Score each candidate process from 1–5 on these factors, then multiply:
- Frequency — How often does this process run? Daily beats weekly beats monthly. A five-minute task done 40 times a week is worth more than an hour-long task done monthly.
- Time per run — Total human minutes consumed, across everyone who touches it.
- Error cost — What does a mistake cost when a human fat-fingers it? Wrong-address shipments and invoice disputes score high.
- Stability — Does the process work the same way every time? Stable, rule-based processes automate cleanly. Processes that change monthly don’t.
Multiply the four scores. Anything above 150 is a strong first candidate. Anything below 40, leave alone — the automation will cost more than it saves.
These thresholds are Westflow planning heuristics, not an industry standard. Check the ranking against your actual process volume, labour cost, error exposure, and implementation quote before approving a build.
The three classic first wins
Three process families are common candidates for a first automation:
- Lead intake and routing. A new lead arrives by form, email, or phone; someone copies it into the CRM, assigns an owner, and sends a first reply. This can be high-frequency, error-sensitive, and stable enough to automate. Measure your current volume and handling time to estimate the likely benefit, then use the lead capture automation guide to plan validation, storage, notifications, and follow-up.
- Invoice and billing handoffs. Data re-keyed from your operations system into your accounting software. Boring, constant, and error-prone: perfect.
- Weekly reporting. If someone assembles the same numbers into the same spreadsheet every Monday, that’s a pipeline, not a job.
What not to automate first
Resist the temptation to start with your gnarliest problem. The worst first projects share a profile: rare, high-stakes, and full of judgment calls. Quarterly planning, custom client proposals, anything involving negotiation — these need humans, and automating around their edges delivers weak ROI while carrying the highest risk of visible failure.
Start boring. Boring is where the money is.
Run the numbers on a real example
Suppose an ops coordinator spends 45 minutes a day moving orders from email into a fulfillment system:
45 min/day × 5 days × 48 weeks = 180 hours a year. At an illustrative loaded cost of $40/hour, that’s $7,200 before counting errors. Compare that baseline with an actual implementation quote, maintenance cost, and expected error reduction to calculate a realistic payback period.
That’s the math to run for every candidate on your list. If you’d like a second pair of eyes on it, use our automation ROI formula and worked example, or request a free automation audit. We score your processes with you and hand you the ranked list, whether or not you hire us to build.