Investigations end with more possible fixes than anyone can do. Without a method, the loudest person or the most senior department picks. An impact/effort grid and multi-voting narrow the list in a way people accept.
Where it comes from
Multi-voting is a standard facilitation method: each person gets a fixed number of votes, agreed before voting and fewer than the number of options, to spread across the list, and the top-voted options go forward. The nominal group technique, from Delbecq and Van de Ven (1971), adds silent individual idea writing first so the first speaker doesn't anchor everyone. The impact/effort grid places each option by how much it helps against how much it takes. Lean Six Sigma's PICK chart, developed at Lockheed Martin, is a version of it, named for its quadrants: Possible, Implement, Challenge and Kill.
| Low effort | High effort | |
|---|---|---|
| High impact | Do first (Implement) | Plan it (Challenge) |
| Low impact | Do if easy (Possible) | Drop (Kill) |
How a bot helps
A bot prepares the options so the vote is informed: each fix in one line, the problem it addresses, the evidence that the problem is real (counts, costs from cost of poor quality), and a rough effort with its basis. It merges duplicates and keeps the original wording visible.
Then people vote. The bot tallies, shows the spread, and flags where departments voted very differently. That disagreement is often the most useful thing on the page.
On the bots' own work
A month of friction notes from a bot team makes a long list of possible skill changes. Grid them by how many corrections each would prevent against how much each takes to build and test.
When to use it
When the list is longer than the team's capacity, and the options are similar in kind. For a few options with real trade-offs, use a decision matrix instead. Light.
Where bots differ
A bot doesn't vote. Its impact and effort placements are drafts with a stated basis, and people move them. A bot's yes isn't a yes.
Sources: Nominal group technique; PICK chart.
Back to Engineering tools for bots.
Short outline
Impact/effort and multi-voting: short outline
What it's for
An impact/effort grid places each possible fix by how much it helps against how much it takes, and multi-voting lets people narrow a long list fairly. A bot prepares the options with evidence and tallies. People vote and decide.
Columns
| Column | What goes in it |
|---|---|
| Option | One possible fix, in one line |
| Problem it addresses | The finding it comes from |
| Evidence | Counts, costs or records that show the problem is real |
| Impact | High or low, with its basis |
| Effort | High or low, with its basis |
| Quadrant | Do first, plan it, do if easy, or drop |
| Votes | Votes received, and from which departments |
| Decision | Go, later or no, and who decided |
Steps, each with its "done when"
- List and merge the options. Done when each option is one line, duplicates are merged, and the original wording is kept.
- Attach evidence and place each on the grid. Done when each option has evidence, an impact and an effort, each with a basis, and people have adjusted the draft placements.
- Vote. Done when each person has cast the agreed number of votes, without seeing others' votes first.
- Tally and show the spread. Done when votes are tallied by option and department, and big disagreements are flagged.
- Decide and record. Done when each option is go, later or no, with who decided and an owner for each go.
Common mistakes
- Voting before the options have evidence.
- Open voting where the first or most senior voter anchors the rest.
- A bot's placements accepted as the decision.
- Using a vote for a few options with real trade-offs, where a decision matrix fits better.