How we score an automation opportunity
The five weighted criteria we use to rank what to build first, why the weights are what they are, and a worked example.
A diagnostic usually turns up 10 to 20 things worth building. A team can take on two or three at a time. The order matters more than the list, so we score every opportunity the same way and let the ranking come out of the scores.
This note sets out the model. It is the same one we use on client work.
The five criteria
Each opportunity gets a score from 1 to 5 on five criteria. Each criterion carries a fixed weight.
| Criterion | Weight | What a 5 looks like |
|---|---|---|
| Financial impact | 35% | Many hours a week, on work that is clearly measurable |
| Ease of implementation | 20% | Fits the tools already in place, few exceptions |
| Data and integration readiness | 20% | Clean data, and the systems involved can be connected |
| Confidence in the outcome | 15% | We have a baseline and the process is stable |
| Risk and oversight | 10% | A mistake is cheap and easy to catch |
The weighted score is each rating multiplied by its weight, added up. The top of the range is 5.0.
Why impact is only 35%
It is tempting to rank by impact alone. The result is a roadmap that opens with the largest and hardest project, which is the one most likely to stall.
Impact gets the largest single weight because the point of the work is to remove hours. The other four criteria together outweigh it because they decide whether those hours actually come off. An opportunity that scores 5 on impact and 1 on data readiness is a data cleanup project with a build attached.
Why risk is only 10%
Risk has the smallest weight because we handle most of it outside the score. Anything that touches money, payroll, sensitive information or messages to customers gets an approval step before it acts, whatever it scored. The weight covers what is left: how much review the system will need once it is live.
A worked example
The numbers below are made up to show the arithmetic. They are not from a client.
| Opportunity | Impact | Ease | Data | Confidence | Risk | Score |
|---|---|---|---|---|---|---|
| Lead routing into the CRM | 4 | 5 | 4 | 4 | 5 | 4.30 |
| Quote drafting from job notes | 5 | 3 | 3 | 3 | 3 | 3.70 |
| Collections reminders | 3 | 4 | 5 | 4 | 3 | 3.75 |
Quote drafting has the highest impact and ranks last of the three. Lead routing goes first: it is nearly as valuable, it is much easier to ship, and it gives the team a live system in the first month. Quote drafting moves up once the job notes it depends on are being captured consistently.
What the score does not do
The score ranks. It does not decide. Two things regularly override it:
- Dependencies. If one build produces the data another one needs, it goes first whatever it scored.
- The owner on your side. A system needs a named person to test it and sign it off. If that person is not available this month, the build waits.
We show the scores, the weights and the overrides in the roadmap, so the order can be argued with.