Solution Prioritization Matrix: Turn a Diagnosed Cause Into a Change Worth Piloting

The Solution Prioritization Matrix sits in the Improve phase of RDMAICS. Its job is narrow and useful. It prioritises feasible solutions by impact, ease and confidence so you can sequence the improvement roadmap.

By the time you reach Improve, you have already done the hard diagnostic work. Measure gave you a baseline. Analyse gave you a cause you can defend. What you have next is a list of ideas. Some are large. Some are quick. Some rest on assumptions nobody has tested. Without a method, the loudest idea wins.

The matrix forces a comparison. Every candidate is scored on the same three dimensions, on the same one to ten scale, by the same group. The tool multiplies the three scores into a Priority Score and ranks the list. You are left with an order, not a debate.

Use it after Analyse, before you commit resources to a pilot. Use it again if new options appear partway through delivery.

There are three moments when this tool earns its place.

The first is when you have more ideas than capacity. Five solutions and one team means four of them wait. The question is which one goes first, not which ones are good.

The second is when the group disagrees. Scoring separates the disagreement into parts. People often agree on impact and split on confidence. That is a far more productive argument than a general one.

The third is when a solution looks obvious. High impact ideas tend to attract attention. The matrix asks whether they are also easy and whether you are confident they will work. Sometimes the obvious answer is still the right answer. Sometimes it is not the right first answer.

The tool ships with sample data. This case is fictional, and it is used here only to show the mechanics of the tool. No real organisation is described.

The scenario is an accounts payable process. Analyse has pointed at delays and rework in how invoices move through the system. Five solutions are on the table.

They are auto-routing with delegation, automated three-way match rules, OCR-based invoice capture, vendor master cleanup, and a vendor portal with validation.

Each one is plausible. Each one would need money and time. The team cannot do all five in parallel. That is the condition the matrix was built for.

The worksheet takes one row per solution and three scores per row. Impact is rated one to ten. Ease is rated one to ten. Confidence is rated one to ten. The tool calculates the Priority Score and the Sequence from those inputs.

In the sample, auto-routing with delegation scores 8 for impact, 9 for ease and 8 for confidence. Automated three-way match rules score 7, 8 and 7. OCR-based invoice capture scores 9, 6 and 7. Vendor master cleanup scores 5, 9 and 8. Vendor portal with validation scores 6, 6 and 5.

Notice what the scoring does to the OCR option. It carries the highest impact score in the set at 9. Nothing else reaches 9 on impact. But its ease is 6 and its confidence is 7. The matrix does not let a single high score carry the row.

Score the three dimensions separately and deliberately. Do impact for all five rows first. Then ease for all five. Then confidence for all five. Scoring across a row invites you to make the numbers agree with a conclusion you have already reached.

The tool multiplies the three scores and ranks the list. The result view shows the ranking as a bar chart on a scale running from 0 to 600, with gridlines at 150, 300 and 450.

Auto-routing with delegation takes sequence 1 with a Priority Score of 576. Automated three-way match rules take sequence 2 at 392. OCR-based invoice capture takes sequence 3 at 378. Vendor master cleanup takes sequence 4 at 360. Vendor portal with validation takes sequence 5 at 180.

Two things stand out. The first is the gap at the top. 576 against 392 is a wide margin. The sequence 1 option is not marginally ahead, it is clearly ahead. That is the kind of result that settles an argument.

The second is the cluster in the middle. 392, 378 and 360 sit close together. Those three are effectively tied. The order between them is produced by small scoring differences that could move if one person changed one rating. Treat the top rank as a decision and the middle cluster as a group.

The bottom of the list is also clear. The vendor portal option scores 180, under half the score of the option above it. Its confidence rating of 5 is the lowest in the set. That is a signal about what you know, not about whether the idea has merit.

The tool also offers an AI review that explains the result in plain language.

The first decision is what to pilot. Auto-routing with delegation at 576 is the clear candidate. High impact at 8, high ease at 9, high confidence at 8. There is no dimension holding it back.

The second decision is what to do with high impact options that are not first. OCR-based invoice capture has the highest impact score in the set at 9 and still lands at sequence 3. The matrix is not telling you to drop it. It is telling you that its ease of 6 and confidence of 7 make it a poor first move. Run the sequence 1 pilot, learn something, then rescore.

The third decision is where to spend investigation time. The vendor portal option scores 5 on confidence. That is the weakest number anywhere in the sample. The way to raise it is not to argue. It is to go and find out. A small test, a conversation with the people who would use it, a look at comparable work elsewhere.

The fourth decision is what to tell a sponsor. A sequence of 1 through 5 with scores attached is a defensible answer to the question of why this and not that.

The most common mistake is scoring ideas that are not solutions. A matrix row should be something you could actually fund. Themes and aspirations will not score honestly.

The second mistake is inflating confidence. Confidence is a statement about evidence, not optimism. If the sample had rated the vendor portal at 8 rather than 5, its score would have risen and the ranking would have shifted. The low rating is the value of the exercise.

The third mistake is reading small gaps as real. The 392, 378 and 360 cluster covers three solutions within 32 points. That is noise, not signal. Do not build a delivery plan on a 14 point difference.

The fourth mistake is scoring alone. One person's ratings are one person's view of the process. Score with the people who do the work and with the people who would pay for the change.

The matrix ranks what you give it. It cannot tell you that a sixth solution exists. If the right answer was never written into a row, no scoring method will surface it. Idea generation is a separate job and it happens before this tool.

The scores are judgements. A Priority Score of 576 looks precise. It is the product of three opinions on a ten point scale. The number inherits every bias in the room. Multiplication makes differences look larger than the underlying confidence in them warrants.

The tool does not size the financial gain. There are no dollars anywhere in this worksheet. Impact of 8 means the group thinks the effect is large. It does not mean a cost has been calculated. If you need a business case, build one separately.

It also says nothing about dependencies. The sequence does not know whether vendor master cleanup needs to happen before automated matching will work. You still need to check the order against the mechanics of the process.

Finally, the matrix does not validate anything. Sequence 1 is a candidate for a pilot, not a proven answer. The pilot produces evidence. Then you come back, rescore, and move on.