Customer needs analysis: find the job behind the request
A step-by-step customer needs analysis: separate requests from jobs, write outcome statements, score them by importance and satisfaction, and use it in sales calls.
A customer needs analysis is the work of finding out what customers are trying to get done, writing each need as a measurable outcome, and ranking those outcomes so you know what to build or sell first. The trap in most customer needs analysis is treating the customer's request ("add a dashboard") as the need, when it is usually one guess at a solution to a deeper job.
So the method below separates what people ask for from the job underneath, borrows the outcome statement format from Tony Ulwick's Outcome-Driven Innovation, and uses importance and satisfaction scores plus the Kano model to prioritize. The last sections show a template table, a made-up worked example and a short version for sales discovery calls.
1. Write down what customers ask for, then treat every request as a clue
Start by collecting the stated needs: feature requests, support tickets, sales call notes, survey comments. Keep the customer's exact words, because you will need them later.
Then stop trusting them as a spec. In a 2002 Harvard Business Review article, Anthony Ulwick argued that customers tend to describe the solutions they want, and that companies which build those requested solutions often see the products fail. His examples were requests like healthier fast food and bigger car engines.
A request tells you where the pain is. It rarely tells you what would remove it. (The customer who asks for a bigger inbox usually wants fewer emails.)
Short version: log every request verbatim, then ask "what were they trying to do when they asked for this?"
2. Name the job before you list a single need
The job is the progress a person is trying to make in a particular situation. The Christensen Institute describes Jobs to Be Done as a lens on the circumstances that push people toward and away from decisions: people "hire" a product when a job comes up in their life.
One of its examples is the milkshake. Morning commuters hired the shake to stave off hunger, make a dull drive more interesting and have something easy to hold in the car, so its real competitors were bananas and bagels. In the afternoon the same shake was a treat for kids, which is a different job with different needs.
The same page says every job has functional, social and emotional parts. Buying a home for a growing family is about space, but also about the neighbors and about feeling like you are taking care of people.
Write the job in the form Strategyn's template uses: a verb, the object of the verb and a contextual clarifier. Its example is "listen to music while commuting to work". Keep products out of the sentence. "Use our app" is not a job.
3. Gather needs by interviewing people and watching them do the job
Interview the people who actually do the job, and watch them do it when you can. Ulwick's company lays out the sequence on its customer needs page: interview customers to break the job into steps, observe them in the setting where the job happens, then use interviews to find the measures they use to judge success at each step, and fill the gaps with more interviews or literature searches.
Strategyn's template breaks the job into eight steps: define, locate, prepare, confirm, execute, monitor, modify and conclude. Walking an interviewee through those steps in order is a good way to stop them from jumping straight to their favorite complaint.
Use one-on-one interviews as your default, so one confident voice cannot steer the room. Whether interviews or focus groups work better is a question Abbie Griffin and John Hauser studied in their 1993 Marketing Science paper "The Voice of the Customer", along with how many customers to interview, how many analysts should read the transcripts and how many needs get missed. It is worth reading before you plan a large study.
Ask about the last time they did the job, not what they would want in theory. Our guides to customer interview questions and The Mom Test cover how to keep people talking about past behavior. For observation, contextual inquiry is the method to borrow.
4. Rewrite each need as a solution-free outcome statement
Turn every raw note into a desired outcome statement. Strategyn's format is a direction of improvement, a metric, an object of control and a contextual clarifier. Its example: "Minimize the time it takes to identify the correct drill bit size for the material being drilled."
The format feels stiff, and that is the point. "Minimize" plus "time" or "likelihood" forces you to say what better means, and leaving out any product makes the statement survive the next redesign.
Expect a long list. The same template says a thorough study typically captures 50 to 150 desired outcomes. If you have eight, you probably wrote features.
Griffin and Hauser describe structuring needs into a hierarchy of primary, secondary and tertiary needs. In practice: group your outcomes under the job steps from section 3, and the hierarchy mostly builds itself. An affinity map helps when the notes are messy.
5. Prioritize with importance and satisfaction, then sort the winners with Kano
Survey a larger group of customers on two questions per outcome: how important it is, and how satisfied they are with their current ability to achieve it. Strategyn uses a ten-point scale for both.
Then score each outcome with Ulwick's opportunity formula, as given on the same template page:
Opportunity = Importance + max(Importance minus Satisfaction, 0)
The max term means a well-satisfied outcome never scores below its importance, and an unmet important outcome counts its gap on top. Outcomes that are important and poorly satisfied rise to the top. Those are the underserved needs.
Do not rank by how often a need came up in interviews. Whether frequency of mention can stand in for importance is a question Griffin and Hauser set out to test, so ask about importance directly instead of assuming it. The loudest customer has one vote.
Once you have a short list of underserved outcomes and some candidate features for them, run a Kano survey on the features. Kano tells you which features customers expect as basics and which would delight them, so you build the must-haves first. Opportunity scores tell you which needs matter; Kano tells you how customers will react to a specific fix.
6. Use this template table to keep the analysis honest
One row per outcome. The first two columns keep the link back to what customers said, so nobody can quietly drop an awkward request.
| Column | What goes in it |
|---|---|
| Stated request | The customer's words, verbatim |
| Source | Interview, ticket, call note, with a date |
| Job step | Which of the eight steps it belongs to |
| Desired outcome | Direction + metric + object of control + context |
| Importance (1 to 10) | Survey average |
| Satisfaction (1 to 10) | Survey average |
| Opportunity score | Importance + max(Importance minus Satisfaction, 0) |
| Candidate fix | A feature, process or offer, written last |
| Kano category | From a later Kano survey on the candidate fix |
Fill the columns left to right. Writing the candidate fix before the outcome is how a needs analysis turns back into a feature request list.
7. Worked example: a dental group that asked for a no-show dashboard
Here is a made-up example. Priya runs operations for a dental group with 12 locations, and she asks a scheduling software vendor for "a dashboard of no-shows by location". She has a spreadsheet tab named NOSHOWS_FINAL_v3, which is how you know she means it.
The vendor's product manager interviews Priya, four front desk leads and two hygienists, and watches a front desk at 7:30 on a Monday morning. The job turns out to be "keep every chair booked through the day". The dashboard was one guess at helping with the monitor step.
The team writes outcomes, surveys front desk and operations staff across its customer base, and scores them:
| Desired outcome | Importance | Satisfaction | Opportunity |
|---|---|---|---|
| Minimize the time it takes to fill a same-day cancellation | 8.6 | 3.1 | 14.1 |
| Minimize the time it takes to see tomorrow's open slots across all locations | 7.4 | 4.0 | 10.8 |
| Minimize the likelihood that a patient forgets an appointment | 8.2 | 6.9 | 9.5 |
| Minimize the likelihood that a hygienist sits idle between patients | 6.1 | 5.8 | 6.4 |
| Increase the likelihood that every location follows the same reminder process | 5.2 | 6.5 | 5.2 |
Check one row: 8.6 + (8.6 minus 3.1) = 8.6 + 5.5 = 14.1. On the last row satisfaction is higher than importance, so the gap counts as zero and the score equals importance.
The dashboard Priya asked for mostly helps with forgotten appointments, which the current reminder texts already handle fairly well. The biggest gap is filling same-day cancellations. So the team prototypes a waitlist that offers the open slot to patients who asked for an earlier time, and runs a Kano survey on that and the dashboard. Priya still gets a simple no-show report, because it is cheap, but it is no longer the roadmap.
8. In a sales discovery call, run the same analysis in 30 minutes
Sales teams use "needs analysis" for the part of a discovery call where you learn what the buyer is trying to fix before you pitch. The steps compress, and the order is the same:
- Ask what prompted the call now. That is the circumstance, and it tells you the job.
- Ask them to walk you through the last time the job went badly. Listen for steps and for the measure they care about (hours, missed revenue, a deadline).
- Restate the need as an outcome in their words: "So the goal is to cut the time from cancellation to refilled slot, mostly on Mondays." Let them correct you.
- Ask how they handle it today and how well that works. That is your satisfaction score without the survey.
- Ask who else feels this and who signs off. Needs differ by role, and the person on the call is rarely the only one.
Only then show the part of your product that addresses the top outcome. Our lists of sales discovery questions and the guide to running a discovery call go deeper on wording. If you coach a team on this, B2B sales coaches are worth a short conversation.
9. Skip the full method when the decision is small or the job is already obvious
The full version, with dozens of interviews, a 100-outcome survey and a Kano follow-up, is for decisions that shape a product or a market entry. It does not fit a single bug fix, a market where you cannot reach enough customers to survey, or a brand-new category where people have no current way of doing the job to rate their satisfaction with.
For a small decision, do sections 1, 2 and 4 on a whiteboard and skip the survey. For a new category, lean on interviews and observation, and treat any scores as rough.
10. Interview the people who do the job every day
The analysis is only as good as who you interview. Interview the person who does the job daily, plus whoever buys and whoever approves, because their outcomes differ. If the right people are outside your customer base, for example CIOs in dental services or experienced user research experts who can sanity-check your outcome list, you can invite specific people through Instant Expert to a short call or a written or voice answer, and pay only if someone books or answers.
Today, pick one customer request from your backlog and write the job and three outcome statements behind it.