Problem Solving in 6 Steps, Including a Brainstorm Session
Updated November 2026

The six basic steps of problem solving are: (1) define the problem as a measurable gap between what is happening and what should be happening, (2) brainstorm possible solutions without judging them, (3) analyse the three or four strongest options for cost, effort and risk, (4) choose one using explicit criteria, (5) turn the choice into a dated, owned action plan, and (6) evaluate the result against the original definition and revise. The order matters: most failed fixes trace back to a problem that was never defined properly in step 1.
This guide is written for a small-business owner or team lead facing a concrete operational problem: late deliveries, a sales dip, a process that keeps breaking. Each step below gives the method, the evidence for why it works, and a running example so you can copy the shape rather than the theory.
Where do the six steps of problem solving come from?
The six-step plan is a practical merge of three older frameworks, each of which is still taught and used on its own. John Dewey described reflective thinking in How We Think (1910) as “five logically distinct steps: (i) a felt difficulty; (ii) its location and definition; (iii) suggestion of possible solution; (iv) development by reasoning of the bearings of the suggestion; (v) further observation and experiment leading to its acceptance or rejection” (Dewey, 1910, chapter VI). George Pólya’s How to Solve It (1945) compressed the mathematical version to four steps: understand the problem, devise a plan, carry out the plan, look back (SUNY Brockport summary of Pólya’s process). The W. Edwards Deming Institute’s Plan-Do-Study-Act cycle supplies the loop that turns a one-off fix into a repeatable process (Deming Institute, PDSA).
The table below is our mapping of the six steps onto those three sources. Use it when a colleague insists their framework is different; usually it is the same steps with different labels.
| Six-step plan | Dewey (1910) | Pólya (1945) | Deming PDSA |
|---|---|---|---|
| 1. Define the problem | Felt difficulty; its location and definition | Understand the problem | Plan (define the change and predict the result) |
| 2. Brainstorm solutions | Suggestion of possible solution | Devise a plan (heuristics list) | Plan |
| 3. Analyse the options | Development by reasoning of the bearings of the suggestion | Devise a plan | Plan |
| 4. Choose a solution | — | Devise a plan | Plan |
| 5. Create an action plan | — | Carry out the plan | Do |
| 6. Evaluate and revise | Further observation and experiment leading to acceptance or rejection | Look back | Study, then Act |
Two things stand out from the mapping. Dewey and Pólya both spend more words on defining the problem than on solving it, and neither one has a separate “choose” step, because a well-defined problem usually narrows the options on its own. The six-step version adds explicit choosing and planning steps because business problems, unlike maths problems, are solved by other people doing things on particular dates.
Step 1: Define the problem as a gap you can measure
Dewey’s line is the whole of step 1: “A question well put is half answered; i.e. a difficulty clearly apprehended is likely to suggest its own solution, while a vague and miscellaneous perception of the problem leads to groping and fumbling” (How We Think, chapter VI). Pólya’s first questions for understanding a problem are the ones to ask yourself: can you state it in your own words, what are you trying to find, what are the unknowns, and what information is missing (Brockport).
A usable problem definition has four parts. Write all four before you let anyone suggest a fix.
| Part | Question it answers | Running example: online orders arriving late |
|---|---|---|
| Current state | What is happening, with a number | In August, 61 of 340 online orders were delivered after the promised date |
| Desired state | What should be happening, with a number | Fewer than 10 late orders a month by the end of November |
| Evidence | Where the numbers came from | Courier tracking export plus customer complaint log |
| Constraints | What you cannot change | No new staff before Christmas; same courier contract until March |
The current-state number is the one most people skip. “Deliveries are slow” is a complaint. “61 of 340 orders were late in August” is a problem, and it also becomes the yardstick for step 6. If you do not have the number, getting it is the first task, not an optional extra.
Two traps at this stage. The first is defining the problem as its assumed cause (“we need a faster courier” is a solution wearing a problem’s clothes). The second is defining it so broadly that no single action could close the gap (“customer experience is poor”). Break a broad problem into two or three narrower ones and run each through the six steps separately. HuddleIQ’s problem-solving templates include a Gap Analysis board (current state against expected result) and a Customer Problem Statement board; either one puts the definition at the top of the canvas where everyone can see it during the later steps.
Step 2: Brainstorm solutions, but not the way Osborn described
Alex Osborn’s original brainstorming rules from Applied Imagination are still the ones most teams use: more ideas is better, wilder is better, build on others’ ideas, and defer criticism. Osborn also claimed that “the average person can think up twice as many ideas when working with a group than when working alone.” That claim has been tested for nearly seventy years and it is wrong. Taylor, Berry and Block found in 1958 that four people brainstorming alone produced nearly twice as many distinct ideas as the same four people brainstorming together, and Michael Diehl and Wolfgang Stroebe’s four experiments, published in the Journal of Personality and Social Psychology in 1987, found that “production blocking” accounted for most of the loss: only one person can talk at a time, so everyone else is holding an idea in their head instead of generating the next one (Diehl & Stroebe, 1987).
The fix is not to abandon group brainstorming but to change its shape. Paul Paulus and colleagues, writing in Frontiers in Psychology in 2018, report that groups of four tend to generate only about half as many ideas as the same number of individuals working alone, that a written idea-exchange procedure (“brainwriting”) yielded twice as many ideas as verbal brainstorming in a short workshop session, and that a solo session after the group session captures ideas the group exchange triggered but nobody had time to voice (Paulus, Baruah & Kenworthy, 2018).
That research translates into a three-phase protocol we use on a shared board:
- Solo first (10 minutes, silent). Everyone writes ideas onto their own sticky notes on a shared HuddleIQ board. Nobody talks, so nobody blocks anyone. Ideas go up as they occur, which means people see each other’s notes and build on them without waiting for a turn.
- Group second (15 minutes). Cluster the notes by theme, read them aloud, and add the combinations and variations that appear. This is where Osborn’s “build on others’ ideas” rule earns its keep.
- Solo again (5 minutes, or overnight). Each person adds anything the discussion triggered. On an online board this can be asynchronous, which suits remote and part-time staff who were not in the room.
For the running example, the solo phase produced 31 notes in ten minutes from four people, including “move the daily courier cut-off from 3 pm to 1 pm”, “batch orders by suburb”, “text the customer a realistic date instead of the promised one”, and “stop selling next-day delivery on Fridays”. HuddleIQ’s brainstorming templates include a Brainwriting board built for the silent written phase, and our guide to running a virtual brainstorming session covers the facilitation side.
Step 3: Analyse your top three or four options
Pick the three or four most promising ideas from step 2 and, for each, answer the same five questions in writing: what would it cost, how long would it take to implement, what would it need from people you do not control, what could go wrong, and how much of the gap in step 1 would it close if it worked. Three or four is the useful number: fewer and you have skipped the comparison; more and the analysis takes longer than the fix.
Before you compare solutions, check you understand the cause. A fishbone diagram sorts possible causes into categories (people, process, equipment, materials, environment) so the team can see whether late deliveries are a courier problem, a picking problem or a promise problem. In the running example the tracking data showed most late orders were placed after 2 pm and missed that day’s pickup, which is a cut-off problem, and it eliminated “change courier” before anyone spent a week costing it.
Should you run a pre-mortem on each option?
Yes, and it takes ten minutes. A pre-mortem asks the team to assume the option has already been implemented and has failed, then to write down every reason why. Gary Klein described the method in the Harvard Business Review in September 2007, citing 1989 research by Deborah Mitchell, Jay Russo and Nancy Pennington which found that this kind of “prospective hindsight” (imagining an event has already happened) increased people’s ability to correctly identify reasons for future outcomes by 30% (Klein, 2007). Its practical value is that it gives the quiet person with the real objection a legitimate way to raise it. For “move the cut-off to 1 pm”, the pre-mortem surfaced that afternoon orders would then be promised for the next day, so the website copy and the confirmation email both had to change, or the same customers would still feel let down.
How do you choose between several good solutions?
Score each shortlisted option against the criteria that matter for this problem, weight the criteria, and pick the highest total, then sanity-check the winner against your gut. The scoring is not there to make the decision for you; it is there to make the decision visible so the team can argue about the weights instead of about personalities.
A weighted decision matrix for the running example, scored 1 (poor) to 5 (good):
| Criterion (weight) | Move cut-off to 1 pm | Batch by suburb | Realistic-date texts | Stop Friday next-day |
|---|---|---|---|---|
| Closes the gap (×3) | 4 | 3 | 2 | 2 |
| Cost (×2) | 5 | 3 | 4 | 5 |
| Time to implement (×2) | 5 | 2 | 3 | 5 |
| Risk to sales (×2) | 3 | 4 | 5 | 2 |
| Weighted total | 38 | 27 | 30 | 30 |
Two rules keep the matrix honest. Decide the weights before you score any option, or people will tune the weights to suit their favourite. And if the winner surprises everyone, do not overrule it silently; find the criterion that is missing from the table, add it, and re-score. In this case “closes the gap” was weighted highest because step 1 had defined the gap in numbers, and the cut-off change won because most late orders were afternoon orders.
Step 5: Turn the choice into a plan someone can follow
A solution is a sentence; a plan is a list of dated, owned tasks. Break the chosen solution into the smallest steps that can each be verified as done, give every step one owner and one date, and put the whole list where the team looks daily. In the running example the plan was seven lines: update the courier pickup time, change the website delivery promise, rewrite the confirmation email, brief the two packing staff, add a 12:45 pm “last orders” check to the daily routine, export tracking data every Friday, and review on 3 October.
There is good evidence for writing plans as if-then statements rather than to-do items. Peter Gollwitzer and Paschal Sheeran define an implementation intention as an if-then plan that links a specific situational cue to a specific response: if situation Y is encountered, then I will do Z to reach goal X (Gollwitzer & Sheeran, overview). Their 2006 meta-analysis in Advances in Experimental Social Psychology found that across 94 independent tests, implementation intentions had a medium-to-large positive effect on goal attainment, d = 0.65 (Gollwitzer & Sheeran, 2006). The mechanism is simple: linking the action to a specific cue means you do not have to remember to decide. “At 12:45 pm each day, the packer checks the order screen and pulls anything that can still make the 1 pm pickup” will happen; “make sure afternoon orders are handled” will not.
Pólya’s advice for carrying out a plan applies here too: check each step as you go and keep an accurate record (Brockport). On a shared board, that record is the task list itself, with each card moved to done by its owner on the day it is done.
Step 6: Evaluate against the original definition, then revise
Step 6 is where the number from step 1 comes back. On the review date, compare the current-state figure with the target, not with a general feeling that things are better. In the running example, September’s export showed 14 late orders out of 355, against a target of fewer than 10. That is a large improvement and still a miss, and both facts matter.
Deming’s version of this step is worth adopting precisely. The Deming Institute describes PDSA as “a systematic process for gaining valuable learning and knowledge for the continual improvement of a product, process, or service”, and explains that Deming insisted on Study rather than Check because checking asks only whether the change succeeded or failed, while studying asks whether the result matched the prediction and what that says about your theory of the problem (Deming Institute, PDSA). In the example, the prediction had been that the cut-off change would remove nearly all late orders. Studying the 14 that remained showed 9 were regional addresses the courier only serviced three days a week, a cause nobody had listed in step 3. That is a second problem, so it went back to step 1 with its own definition.
If the plan missed by a wide margin, the options are: revise the plan (more steps, different owners, an earlier cut-off), return to the step 3 shortlist and implement the runner-up, or, if the evaluation showed the cause was not what you thought, redefine the problem. Pólya’s “look back” adds one more question: can the method be reused? A cut-off check that fixed late deliveries can probably fix late click-and-collect orders too.
How long should each step take?
For a problem the size of the running example, defining takes an hour (most of it finding the number), brainstorming takes half an hour using the three-phase protocol, analysis and choice take one meeting of about an hour, planning takes thirty minutes, and evaluation takes fifteen minutes on the review date. The whole cycle is under four hours of meeting time spread over a month. Problems that need more than a day of meetings are usually two or three problems bundled together; split them in step 1.
The step people underspend on is the first and the step they overspend on is the third. Analysis paralysis is common because analysing feels productive and risk-free, whereas choosing means owning an outcome. Setting a time-box for step 3 (one meeting) and using the decision matrix to force step 4 are the two habits that shorten the cycle most.
Does the six-step plan work for teams, or only for individuals?
It works for both, but the team version is a different skill and needs the tools above, not just the steps. The OECD’s PISA 2015 assessment measured collaborative problem solving separately from individual problem solving across 51 education systems, and found that on average across OECD countries only 8 percent of 15-year-olds reached the top proficiency level, while 28 percent could solve only straightforward collaborative problems, if any (NCES, PISA 2015 Collaborative Problem Solving). PISA’s definition of the skill is the useful part: solving a problem “by sharing the understanding and effort required to come to a solution, and pooling their knowledge, skills and effort to reach that solution”. Sharing the understanding is step 1 done in public; pooling the effort is steps 2 to 5.
In practice that is why we run the six steps on a single shared board rather than across a meeting, an email thread and a spreadsheet. The problem statement, the brainstorm notes, the fishbone, the decision matrix and the task list sit on one canvas, and a person joining at step 4 can scroll up and read how the team got there. HuddleIQ’s online whiteboard supports this with an infinite canvas, sticky notes, tables and real-time or asynchronous editing; the features page lists the full set.
Frequently asked questions
What is a step-by-step procedure for solving a problem?
Work through the six steps in order: define the problem, brainstorm solutions, analyse the strongest options, choose one, turn it into a plan, and evaluate the result against the original definition. Each step feeds the next, and the first step, a measurable definition of the problem, is the one most often skipped.
What is the first step in problem solving?
Defining the problem as a measurable gap: the current state with a number, the desired state with a number, where the numbers came from, and the constraints you cannot change. Dewey called this “location and definition” of the difficulty and argued that a question well put is half answered. If you cannot write the current-state number, finding it is the first task.
Is brainstorming in a group better than brainstorming alone?
Not in the traditional talking format. Diehl and Stroebe’s 1987 experiments found that people brainstorming alone and pooling their ideas outproduce interacting groups, mainly because only one person can speak at a time (Diehl & Stroebe, 1987). Groups do better when everyone writes ideas simultaneously first, discusses second, and adds a solo round afterwards, which is easy to run on a shared online board.
How do I choose between several solutions?
Use a weighted decision matrix: agree the criteria and their weights before scoring, score each shortlisted option from 1 to 5, and total them. The matrix makes the trade-offs visible so disagreement moves from opinions to weights. If the result surprises the team, look for the missing criterion rather than ignoring the score.
What should I do if my solution does not work?
Compare the result with the step 1 number to see how far off you are, then study why. A near miss usually means revising the plan; a wide miss with a cause you had not listed means the problem was defined wrongly, so go back to step 1. Deming’s PDSA cycle treats this as normal: each cycle tests a prediction and refines your understanding of the problem.
Which problem-solving framework should a business use?
Any of them, as long as it starts with definition and ends with evaluation. Dewey’s five steps, Pólya’s four, Deming’s PDSA and the six-step plan map onto each other almost line for line (see the table at the top of this article). Pick the one your team will follow and keep the artefacts, especially the problem statement and the decision matrix, on a board the whole team can see.
If you want to try the six steps on a live problem this week, open a board from the problem-solving templates, write the four-part problem statement first, and do not let anyone add a solution until the current-state number is on the canvas.
Ready to Collaborate Better?
Try HuddleIQ free and transform how your team works together.
Get Started Free