🎯 Quick Answer
What: A solution selection matrix is a simple grid for choosing between possible fixes. The fixes go in the rows. The questions you judge them by go in the columns. Each fix gets a plain rating on each question, and the comparison shows which fix to try first.
Why it matters: Once a team finds the cause of a problem, it often takes the first fix someone suggests. That fix is usually the one someone wanted before the analysis started. A matrix makes the team list several fixes and compare them on the same terms.
How to apply: Write the root cause at the top. List at least three fixes, and five or more if you can. Agree on four or five questions before you rate anything, such as: Does it remove the cause? How much effort is it? How soon can we try it? What could it break? Rate each fix with the people who will use it. Look hard at the top two, and write down why the others lost.
The payoff: Better fixes, often smaller and faster than the first idea, and a one-page record of why the team chose what it chose.
Many improvement projects go wrong at the same point. The team has done the hard part. It found the cause, and it can prove it. Then, in the next meeting, someone suggests a fix. Everyone nods, and that becomes the plan.
The fix may be fine. But nobody knows, because nothing was compared to it.
This article covers why the first fix wins so often, how to find better ones, and how to use a solution selection matrix to choose between them. It ends with a field story about a fix that almost lost because it looked too small.
Why the first fix usually wins
There are three plain reasons.
It was there first. A fix that someone wanted before the analysis is ready the moment the cause is named. Often it is a purchase, a new system or an extra check that was proposed a year ago and turned down. The analysis now looks like a reason to approve it.
It often comes from the most senior person in the room. When a manager names a fix, most people stop offering others. They are not being timid. They can see which idea the manager wants.
A meeting feels finished once it has an answer. Comparing options takes another hour, and nobody has asked for that hour.
None of this means the first fix is wrong. Sometimes it is the best one. The problem is that without a comparison, there is no way to tell.
Start from the cause
A fix can only be judged against something. That something is the root cause the Analyze phase proved. Write it at the top of the page, in one sentence, before anyone suggests anything.
This matters because many popular fixes do not touch the cause at all. They catch the problem after it happens, or they hide it. An extra inspector at the end of the line catches bad parts. It does not stop them from being made.
If you are not sure the cause is proven, go back a step. The root cause has to be right before any fix can be.
How to find more than one fix
Most causes can be fixed in several different ways. A simple way to find them is to ask four questions about the cause:
Can we change how the work is done? A different method, a different order of steps, or one standard way instead of three.
Can we change the tools or equipment? A guide, a fixture, a gauge, an adjustment or a different machine.
Can we change a rule, a schedule or a setting? Who approves, how often something runs, or what limit triggers an action.
Can we change the material or what comes in? A different supplier, a different specification, or a different way parts are packed.
Each question tends to bring up a different kind of fix. Some are big. Many are small.
A few rules help the list grow:
Collect every idea before judging any of them. When ideas are judged as they are spoken, the list usually stops at the first good one.
Ask the people who do the work. They often know a fix that nobody ever brings to a meeting.
Write down the ideas that sound too small. Small fixes are the ones most often left off the list.
Aim for at least five fixes before anyone says which one they like.
How to build a solution selection matrix
The matrix is a simple grid. The fixes go down the left side. The questions go across the top. Here is how to build one in five steps.
Step 1: Agree on the questions first
Pick the questions before you look at the fixes, so nobody shapes them to suit a favorite. Four or five are enough. These four cover most projects:
Does it remove the root cause?
How much effort will it take, in time, money and people?
How soon can we try it?
What else could it affect?
Some teams add a fifth question, such as "Will it still work in six months?" Use whatever matters most for your project, but keep the list short.
Step 2: Rate each fix on each question
A simple scale is enough: low, medium or high. Some teams use 1, 3 and 9 to spread the scores apart. Either works. The matrix is there to make the comparison visible. It is not there to produce a precise number.
If some questions matter more than others, you can give each one a weight. Be careful with this. Weights set after the rating are easy to adjust until a favorite fix comes out on top. If you use weights, agree on them in Step 1.
Step 3: Rate it with the people who will use the fix
Operators, technicians and the people who run the process every day often know right away why a fix will not work on the night shift or with a different supplier. Ask them before the fix is installed, not after.
Step 4: Look hard at the top two
The highest rating is a starting point, not a final answer. Look at the top two fixes and ask two things about each. What could it break? How could we try it small? A short FMEA on the changed process is a good way to answer the first question. Often the second-best fix becomes part of the winner.
Step 5: Write down why the others lost
One line per fix is enough. When the problem comes back, or a new manager asks why the team did not buy the bigger machine, the answer is already on paper.
A worked example: the crooked labels
A packaging line kept putting labels on crooked. The team proved the cause: each shift loaded the label roll at a different tension.
The team listed five fixes:
A. Buy a new labeler.
B. Add an inspector at the end of the line.
C. Mark the right tension setting on the machine, and teach one standard way to load a roll.
D. Retrain every shift.
E. Run the line slower.
Then they rated each one.
A. New labeler: probably removes the cause; high effort; months to try; a new machine can have new problems.
B. Inspector: does not remove the cause, it only catches crooked labels; medium effort; can start this week; adds a cost every week.
C. Marked setting and one method: removes the cause; low effort; can be tried in one afternoon; one setting may not suit every label supplier.
D. Retrain every shift: partly removes the cause; medium effort; one week; people forget as shifts change.
E. Slower line: does not remove the cause; low effort; can start today; less output every hour.
Two fixes did not touch the cause, so B and E were out. The new labeler might work, but it would take months, so it moved to the back of the list. People would forget the training over time, so the team made it part of C: mark the setting, then teach the method once.
The team tested C on one shift for two weeks, with each label supplier, before using it on every shift. Crooked labels became rare. The new labeler was never ordered.
The fix that looked too small
Here is a field story about a different mistake: picking a fix because it looks big enough.
A fabrication shop had too many welded frames failing inspection. Each one had to be ground out and welded again. Two engineers looked into it and found the cause in the parts, not in the welding. One cutting machine had drifted out of calibration. Its parts came out slightly off, so the gaps were uneven when welders fit the parts together. Uneven gaps led to weld defects.
The team listed four fixes: hire another weld inspector, buy a new cutting machine, recalibrate the old machine and check it every week, or give each welder a simple gauge to check the gap before welding.
The new machine was the fix most people expected. It looked serious. It would look good in a report to management. The team almost proposed it for exactly that reason. The gauge was a bent piece of steel. The recalibration took a morning. Together, they looked too small to present as the answer.
But the matrix showed what each fix actually did. The recalibration removed the cause. The gauge caught any bad part before a welder spent time on it. The new machine would also have removed the cause, but with far more effort, and months later.
They chose the small fixes. Within a few months, rework on those frames had dropped to a fraction of what it was. The new machine was never needed.
Judge a fix by whether it removes the cause, not by how big it looks.
What leaders can do
Leaders affect this step more than any other, mostly through the questions they ask.
When a team brings a fix for approval, ask: "What else did you consider?" It is not a way to say no. It checks whether the team compared options. If the answer is "nothing," send them back for an hour, not a month.
When you have your own idea, share it last. If the most senior person speaks first, their idea becomes the first fix, and the list stops growing.
And when a team chooses a small fix over a big one, support it. Small fixes are easy to dismiss in a budget meeting. They are also the ones most likely to be tried this week.
Choose, then try it small
The matrix picks the fix to try first. It does not prove the fix works. In DMAIC, the step-by-step method used in Lean Six Sigma, the next step is a small trial on one line, one shift or one area, before the change goes everywhere. When you check the trial, compare the results to the process's normal ups and downs, so that one good week is not mistaken for a real change.
If the matrix leaves you with more good fixes than the team can try at once, the same thinking used to prioritize improvement projects applies: try fewer, and finish them.
I write one of these every week, based on what actually happens on projects rather than what the textbook says should happen. You can join the newsletter here.
FAQ
What is a solution selection matrix?
It is a grid for choosing between possible fixes. The fixes go in the rows, the questions you judge them by go in the columns, and each fix gets a rating on each question. It makes the comparison visible, so the team judges every option on the same terms.
What criteria should a solution selection matrix use?
Keep it to four or five. The most common are: does it remove the root cause, how much effort it takes, how soon it can be tried, and what else it could affect. Some teams add cost, risk or how long the fix will last. Agree on the criteria before you rate any fix.
How many solutions should we consider?
At least three, and five or more is better. With one option there is nothing to compare. With only two, the list usually holds the first idea and one alternative nobody liked.
Should the criteria be weighted?
Only if they really differ in importance, and only if the weights are agreed before anyone rates a fix. Weights set after the rating are easy to adjust until a favorite wins.
What comes after choosing a solution?
Try it small first: one line, one shift or one area, long enough to see whether it works. Compare the results to the process's normal range before you use it everywhere. Then make it the standard way of working, so the gain holds.
List more than one fix. Choose the one that removes the cause.
