← Back to blog
·11 min read
Written by:
CL
Casey Lin
Verified by:
MI
Morgan Ito

A Pain Point Scoring Model You Can Run Yourself

A six-factor pain point scoring model with weights, a worksheet, a hypothetical worked example and cut-offs, so you can prioritize pain points by hand.

Share:

Key Takeaways

  • The model scores each problem 1 to 5 on six factors, then weights them so the total runs from 20 to 100.
  • Severity, frequency, recency, buyer intent, willingness to pay and room to compete carry weights of 20, 20, 10, 15, 20 and 15.
  • Two gates apply before any total counts: a severity or willingness-to-pay score of 1 caps the problem at the hold band.
  • A tool can automate severity, frequency, recency window and competitor gaps, but buyer intent and willingness to pay still need your judgement.
  • The weights are a starting default, not a standard, so set them before scoring and keep them fixed across problems.

Quick answer: A pain point scoring model rates each problem from 1 to 5 on six factors (severity, frequency, recency, buyer intent, willingness to pay and room to compete), multiplies each by a weight, and adds them into a total from 20 to 100. Use it to rank problems you have already collected, then test the top one before building.

Several posts here touch on prioritization, so it helps to say what this one is for. Our how to prioritize pain points guide explains the thinking, including the painkiller versus vitamin test. The pain point research template is the spreadsheet where you log raw evidence. The startup idea validation scorecard judges a whole idea, including whether you can reach the audience.

This page sits between them. It scores a single problem, using six factors with fixed weights and written definitions, so two people scoring the same evidence get close to the same number. It also shows which parts a tool can automate, using the pipeline described in how PainPointMap scores pain points.

The Six Factors and Their Weights

Each factor is scored from 1 (weak) to 5 (strong). The weight says how much that factor can contribute to the 100-point total. These weights are a reasoned default, not an industry standard. Change them if you like, but do it before you look at any evidence.

FactorWeightWhat you are asking
Severity20When this happens, how much does it cost people in time, money or stress?
Frequency20How many separate people raise it, and how often?
Recency10Is it still being discussed now, or only years ago?
Buyer intent15Are people asking for a solution, a tool or a recommendation?
Willingness to pay20Is anyone already spending money or serious time on a workaround?
Room to compete15Do current options fail in a specific, nameable way?

The formula for each factor is the score divided by 5, times the weight. Add the six results. A problem that scores 1 everywhere totals 20, and one that scores 5 everywhere totals 100.

Willingness to pay and severity are weighted highest on purpose. Plenty of problems are frequent and trivial. People mention them and never pay.

Scoring Rubric for Each Factor

Write the definitions down before scoring. These bands are working defaults you can adapt.

Factor135
SeverityMild annoyanceSlows work or causes occasional lossBlocks work or costs money regularly
FrequencyOne threadA few separate threadsMany separate threads, several communities
RecencyOnly old postsSome posts in the past yearSeveral posts in recent months
Buyer intentVenting onlySome "any tool for this?" postsFrequent explicit requests for a solution
Willingness to payNo spend mentionedTime cost or a small tool mentionedNamed paid tools, hired help or budgets
Room to competeWell served, few complaintsMixed: adequate tools, clear gripesNamed tools fail the same way repeatedly

Scores of 2 and 4 sit between the bands. When you are unsure between two numbers, pick the lower one and write down why.

Counting rules that keep scores honest

  • Count distinct threads and people, not comments. One long argument is one data point.
  • Log a link and exact quote for every score above 3. A score you cannot point to is a guess.
  • Score room to compete only after you have named at least two existing options. If you cannot name any, score the factor 3 and flag it for research instead of giving a 5.
  • Treat upvotes as attention, not pain. A popular post can be a joke.

If you need a place to collect those links and quotes, the Reddit research checklist and the template above cover it, and how to find pain points on Reddit shows the searches.

The Worksheet

Copy this table into any spreadsheet. One row per problem. The total column is the only formula.

ProblemSeverity (x4)Frequency (x4)Recency (x2)Buyer intent (x3)Willingness to pay (x4)Room to compete (x3)TotalGate hit?
Problem name1 to 51 to 51 to 51 to 51 to 51 to 5Sum of score times multiplierYes or no

The multipliers are the weights divided by 5, which lets you skip the division. For example, a severity of 4 is 4 times 4, or 16 points.

Add two more columns outside the formula: evidence links and biggest doubt. The doubt column is where most weak ideas reveal themselves, and it is worth reading before you look at the total.

Gates Before You Trust the Total

A weighted sum can hide a fatal flaw. A problem with perfect frequency and a severity of 1 might still total a decent number. Two gates prevent that.

  1. Severity gate. If severity is 1, cap the problem at the hold band no matter what the total says.
  2. Willingness-to-pay gate. If willingness to pay is 1 and you cannot find any spend even after searching for it, cap the problem at the hold band.

Gates are a deliberate break from pure arithmetic. They say that some factors are necessary, not just additive. This is the same idea behind the minimum scores in our older prioritization framework.

Worked Example (Hypothetical)

All numbers below are invented for illustration. They are not measured data and not real quotes. Imagine you researched three problems in a freelancer and small-business setting and scored each one.

Problem (hypothetical)SevFreqRecIntentWTPRoomTotal
Freelancers chase late invoices45434275
Podcasters struggle to book guests33543470
Landlords dislike tenant screening forms52323361

The arithmetic for the first row: severity 4 gives 16, frequency 5 gives 20, recency 4 gives 8, buyer intent 3 gives 9, willingness to pay 4 gives 16, and room to compete 2 gives 6. The total is 75.

What the numbers say, and do not say:

  • The invoice problem wins on frequency and willingness to pay, but its room to compete is low. Many tools already exist, so a total of 75 does not mean it is easy. The biggest doubt would be "a contract template solves this better than software."
  • The podcast problem is a close second and has the best room to compete and the best intent. A pilot here may be cheaper to run than a fight against incumbents.
  • The landlord problem has the highest severity but few people raising it. It fails nothing at the gate, yet its total is lowest. It may be a good niche for interviews, not for a broad scan.

Nothing here says which will make money. It says which to investigate first. This is also why the Reddit pain point examples are worth reading side by side with your own scores.

Interpreting the Total

Use bands rather than a single magic number.

TotalAction
70 to 100Test: run interviews or a small paid pilot this week
55 to 69Hold or refine: narrow the audience or gather missing evidence
Below 55Drop for now: log it and revisit if the evidence changes

Differences of a few points are noise. A 72 and a 68 are the same bucket in practice, so choose between them using the doubt column and how easily you can reach the people. If the top two are close, our seven-day Reddit validation plan gives you a fast way to test one without ignoring the other.

How the Model Maps to What a Tool Automates

The manual model is deliberately close to what a Reddit analysis tool produces, which makes it a good way to understand any tool's output. Here is how it maps to PainPointMap, based on how a scan works.

Model factorAutomated by a scan?How
SeverityPartlyThe scan returns a 1 to 100 severity rating produced by an AI model, which you can convert to a 1 to 5 band
FrequencyPartlyThe scan returns the model's count of posts in its sample that mention the problem
RecencyBy settingYou choose a 7 day, 30 day, 90 day or 1 year window, so recency is controlled by your range
Buyer intentNot scoredYour own keywords can steer which posts rank first, but you judge the intent by reading the threads
Willingness to payNot scoredNo payment data is involved. Look for spend in the cited threads
Room to competeProfessional onlyCompetitor mapping and a market gap score give a starting list, generated from model knowledge and worth verifying

Two honest cautions. First, the tool's numbers come from a sample of recent posts, and at most 50 posts are sent to the model, so frequency is a count inside that sample rather than across Reddit. Second, the scores are relative to one scan, so do not compare a 70 from one subreddit and window with a 70 from another.

The practical workflow is to let a scan do the reading and clustering, then bring the top candidates into your worksheet and fill the factors a tool cannot judge. That leaves your time for the part that matters, which is deciding what the evidence means. For the wider trade-offs, see manual vs automated Reddit research and how to choose a Reddit research tool.

Start your 3-day free trial if you want to try it. A card is required, you get 1 scan a day during the trial, and it then bills $19 (Starter) or $49 (Professional) a month unless you cancel.

Limits of Any Scoring Model

Scoring creates a feeling of rigor, so be wary of it. Four limits apply to this model and to tool-based scoring alike.

  • Garbage in, garbage out. A 5 on frequency from three threads in one angry subreddit is still thin evidence. Our post on Reddit data limits lists what Reddit cannot tell you.
  • False precision. A total of 73 looks more exact than a human 1 to 5 rating deserves. Use bands.
  • Weights encode opinion. If you favour B2B, you might raise willingness to pay. Fix your weights first, then score.
  • Anchoring on favourites. People score ideas they love higher. Have someone else score a few rows blind, or score the problem before you think about the solution.

For more on the mistakes that skew results, read Reddit research mistakes and the guide to market gap examples. If your audience is businesses, finding B2B pain points on Reddit shows which phrases signal budget.

Final Thoughts

A scoring model is a forcing function. It makes you write down what you saw, weigh it against a fixed standard, and notice when your excitement is outrunning the evidence. Six factors, three action bands and two gates are enough to rank a shortlist, and the ranking tells you where to spend your next hour, not whether to quit your job.

If you would rather skip the reading, start your 3-day free trial of PainPointMap and bring its output into your worksheet. A card is required, the trial includes 1 scan a day, and afterwards it is $19 (Starter) or $49 (Professional) a month unless you cancel. More guides live in the Reddit research guides hub, and the niche ideas by platform page can help you pick where to look next.

Frequently Asked Questions

What is a pain point scoring model?

It is a repeatable way to turn evidence about a problem into a number you can compare. You rate each problem on a few factors such as severity and frequency, weight the factors by importance, and add them up. The result ranks problems consistently, but it is only as good as the evidence behind each rating.

Which factors should a pain point scoring model include?

At minimum, how often the problem appears and how much it hurts. Add recency so old complaints do not dominate, buyer intent so you favour people actively looking for a fix, willingness to pay, and room to compete. The six-factor model below uses all of these, and you can drop one if you cannot gather evidence for it.

How do I weight pain point scores?

Give more weight to factors that predict a paying customer, typically severity, frequency and willingness to pay. Lower weights suit recency and competition. The weights here sum to 100 so each total reads as a percentage. Decide them before you score anything, otherwise you will be tempted to tune them toward a favourite idea.

Is a high pain point score proof that I should build?

No. A score ranks the problems you studied against each other using the evidence you collected. It cannot tell you whether enough people will pay, and Reddit is a limited sample. Use a high score to choose what to test next, through interviews, a landing page or a paid pilot.

Can a tool score pain points for me?

A tool can handle the slow part, which is reading many posts and producing severity and frequency numbers with source links. It cannot confirm that anyone will pay. Whatever you use, open the cited threads and check them. Scores from any tool are an input to your decision, not the decision.

Find what your customers actually complain about.

Scan the communities where your customers talk and see real pain points ranked by how often they come up — no surveys, no guesswork.

Scan My Market: 3-Day Trial
CL
Casey Lin
Research Writer, PainPointMap

Covers competitor analysis, SaaS go-to-market strategy, and how founders use community research to find product-market fit.