Customer and User Obsession Questions
Reasoning from the customer or end user inward when making decisions in any role: building empathy for users, including buyers who are not the people who use the product; identifying and prioritizing customer pain points and workarounds; collecting and acting on customer feedback, including how representative it is, how it compares with usage data or satisfaction scores, and how it reaches the people who can act; bringing the voice of the customer into roadmap and technical trade-offs; advocating for users against internal or deadline pressure; and balancing customer needs against business goals and engineering constraints, such as a large account's bespoke request, power users versus everyday users, or retiring a feature. Includes stories of changing course because of customers or recovering after letting one down. Assesses whether a candidate starts from the customer's problem rather than from features or technology. Study design and fieldwork, research synthesis, personas and journey maps, analytics and experimentation mechanics, reliability engineering, customer-success operations, market and competitive research methods, and live handling of angry customers are covered elsewhere.
Power users want fine-grained control and everyday users want a simple, safe default. How do you decide what to build so both are served without overcomplicating the product?
Sample Answer
Direct answer
I serve both with a safe default for everyday users and progressive disclosure for power users: the common path is simple and works out of the box, and fine-grained control lives one step deeper (an "Advanced" area, a rule editor), with guardrails. I decide per control, using who needs it, how often, and how safely it can be hidden.
Decision test for each requested control
- Is a real job blocked without it? Ask power users what they are trying to achieve, not which control they want.
- Can the default path stay unchanged? If adding it clutters the main screen, it goes behind a deeper layer.
- Is it safe to misconfigure? If a wrong setting can cause harm, add a preview, a confirmation, and an undo.
- How many people need it? Check usage evidence and who the power users are (revenue, influence, advocacy), but do not let the loudest segment set the default.
- Can a preset do the job? Presets cover most of the need without exposing every knob.
Worked example (illustrative): alert rules in a monitoring product
Everyday users get three presets: "Recommended", "Quiet" and "Everything". Power users get a rule editor behind "Advanced" with a preview of how many alerts the rule would have sent last week and an undo for edits.
| Requested control | Decision |
|---|---|
| Custom thresholds per metric | Advanced editor |
| Quiet hours | Main settings (many need it) |
| Per-channel routing | Advanced editor |
| Raw query syntax | Advanced editor, with preview |
| Turn alerts off completely | Not offered; use "Quiet" with a confirmation |
Five requests: one on the main screen, three behind Advanced, one declined in favour of a safer route. Measure the share of everyday users who keep the default and whether power users complete their jobs without support.
For an AI-based product
Expose a few outcome-level choices (for example a "creative" versus "precise" mode) as the default, and keep model-level settings in an advanced layer.
Pitfalls
- Settings sprawl: every extra control is a decision for every user.
- Hiding power features so deeply that experts leave.
- Treating the loudest segment as the whole user base.
- No safety net on advanced controls.
Customers have built manual workarounds for something the product does not do well. How do you decide whether those workarounds signal a problem worth fixing, and what would you do about it?
Sample Answer
Direct answer
A workaround is customers paying in their own time and effort for something the product does badly, which makes it stronger evidence than a request for a feature. Still, not every workaround deserves a fix. I judge it with six checks, observe it live to learn the real goal behind it, size it with data, and then choose between a native fix, a lighter option, documenting it, or deliberately leaving it.
Six checks
| Check | What to look for |
|---|---|
| Frequency | How many accounts, how often (daily, weekly, quarterly) |
| Cost per occurrence | Time spent, error risk, anything lost on the way |
| Investment | How much customers have put into it: a sticky note, or a spreadsheet with macros someone maintains (heavy investment means a strong need) |
| Criticality | How much hangs on it: does it sit in a core flow, or at the edge of the product |
| Persistence | How long it lasts: still in use months later, taught to new hires |
| Fit | Should the product own it, or is an integration or partner the better home |
A workaround scores as worth fixing when it is frequent, costly, persistent and in a core flow.
What I would do
- Watch three customers perform it (screen share). Ask what the output is for. The goal behind the workaround is often narrower than the workaround itself.
- Size it from data: support notes, usage events, account lists. Treat self-reported time as an estimate and confirm it by observation.
- Choose a response: native fix, a lightweight version (template, export, API access), documentation that makes the workaround official, or no change with a reason.
- Prototype with the customers who built the workaround and check whether they drop it.
Worked example (illustrative numbers)
Finance teams export invoices, re-sort by due date in a spreadsheet and email the list every Monday. Reviewing 40 accounts through support notes and calls, 14 do this weekly, about 45 minutes each (self-reported).
14 accounts x 0.75 hours x 52 weeks = 546 hours per year
That is about 14 forty-hour working weeks of customer time across only the 14 accounts found. Watching three of them shows the real goal: their manager wants a due-date-sorted list on Monday morning. These are different bases: 546 hours is recurring annual time spent by the customers' own staff (self-reported, not a cost to the vendor), while three engineer-weeks is a one-off vendor build cost, so the two are not subtracted from each other. The comparison shows that a small one-off build could remove a large recurring burden for the accounts inspected. Against a build estimate of, say, three engineer-weeks for a saved view plus scheduled email, the case is strong, and a lightweight version solves the stated goal.
Scoring the other four checks for the same workaround (illustrative):
| Check | Evidence | Rating |
|---|---|---|
| Investment | A shared spreadsheet template with a sort macro, passed between finance teams | High |
| Criticality | Feeds the weekly payment run, a core finance task | High |
| Persistence | Same routine reported by accounts that signed up over a year ago, and taught to new hires | High |
| Fit | The data lives in the product, so the product should own it | Native |
Frequency (14 of 40 accounts weekly) and cost (45 minutes) were already assessed above, so all six checks point the same way. A workaround that scored high on one check (investment) and low on the other five (say, one customer with an elaborate macro, seen once, at the edge of the product) would instead go to documentation or a services conversation.
Trade-offs and pitfalls
- You only see workarounds from customers who stayed. Those who gave up and left left no spreadsheet behind.
- Some workarounds reflect an intentional product boundary. If the product is deliberately narrow, the right answer may be an integration.
- A workaround that is elaborate but rare (one customer, one macro) is a services conversation (a paid, customer-specific piece of help from a services or solutions team), not a roadmap item (a feature the product team plans to build for everyone).
- Do not extrapolate the 546 hours to the whole customer base without sampling; the count only covers accounts you inspected.
Your dashboard says a new feature is a success, but interviews and tickets show users are frustrated. How do you work out which signal to trust, and what do you decide?
Sample Answer
Direct answer
I would not pick a winner by default. I would first ask what the dashboard metric actually measures, and whether it could be true at the same time as the frustration. Most often they are both right about different things or different people. Then I would run a short check of the metric against outcomes, decide on a conditional rollout with an explicit stop rule, and fix the cause.
Step 1: Find how both could be true
| Possible reason | Test |
|---|---|
| The metric counts clicks or visits, not success (a proxy is a measurable stand-in for what you really care about; users may be drawn to a new button or forced to use it) | Look at task completion, retries and time to finish |
| The average hides a segment | Split by new versus existing users, plan and platform |
| Novelty spike (heavy use only because the feature is new) then drop | Look at repeat use by cohort (users who started in the same week) over weeks |
| The people who talk are not the people measured | Compare complainers' behaviour with the whole base |
| Tracking bug | Check events against a few real sessions |
| Praise comes from users who stayed, while those who left said nothing | Study the people who dropped, not only the quoted ones |
Worked example (illustrative numbers)
Task completion overall rises from 71% to 72.2%, which looks like success. Existing users are 70% of volume and new users 30%.
- Before: existing 80%, new 50%: 0.7 x 80 + 0.3 x 50 = 71%.
- After: existing 86%, new 40%: 0.7 x 86 + 0.3 x 40 = 72.2%.
Existing users improved by 6 points and new users fell by 10, so the interviews with frustrated newcomers were right and the dashboard was too.
Why the total rose although newcomers got worse: the total is a weighted average, and existing users carry 70% of the weight. Their 6-point gain adds 0.7 x 6 = 4.2 points to the total, the newcomers' 10-point loss subtracts 0.3 x 10 = 3.0, and the net is +1.2, which is 72.2 - 71. The 70/30 mix did not change here; the average simply hides two opposite movements. (Simpson's paradox is the stronger case, where a trend in every group reverses when groups are combined because group sizes differ. This is a cousin of it, not an instance.) That is why I split by segment before trusting an average.
Related cases and how I read them
- Positive quotes but a drop in the funnel (the steps from starting to finishing, such as sign-up to first success): quotes come from those who made it through. Study where the others left.
- Clicks up but users say results are irrelevant: a click shows curiosity, not value. Use what happens next: saves, purchases, returns, "not interested" signals.
- Conversion (the share of users who complete the desired action) up but tickets doubled: convert tickets to a rate. If tickets went from 120 to 240 while active users went from 10,000 to 20,000, the rate is 12 per 1,000 users both times, which is no worse. If users stayed flat, the rate really doubled. Also look at refunds, cancellations and support cost, because a confusing flow can lift conversion short term.
- A good average with a segment complaining: the segment is the finding.
What I decide
Do not declare success. Keep the feature live, hold any wider expansion, add a counter-metric (a second metric watched to catch harm, such as day-14 retention, the share of adopters still active 14 days after first use, or tickets per 1,000 users), and agree a stop rule before looking at next week's numbers.
An example stop rule (illustrative): before launch new-user task completion was 50% and tickets ran 12 per 1,000 active users. After launch, new-user completion is already 40% (the worked example above), so a rule written as 'stop if it drops below 45%' would already be breached and would just say stop. The honest version is therefore anchored to the current reading: expansion stays on hold now. We resume expanding only when new-user completion is back to at least 45% and tickets are at or below 15 per 1,000 in two consecutive weekly readings after the targeted fix. If new-user completion falls below 38% or tickets rise above 18 per 1,000 in either of the next two weekly readings, we pull the feature for new users while we fix it. The thresholds must sit outside the normal week-to-week wobble, which I check against earlier weeks, and I review every Monday for four weeks. Ship a targeted fix for the struggling segment and re-measure.
Which signal wins: the one closer to the user's real outcome and less exposed to selection (a bias where the measured group differs from everyone, for example only those who stayed). Neither the dashboard nor the quote file earns trust on its own.
A stakeholder tells you customers asked for a dense, information-heavy dashboard and wants it built exactly as requested. What you have seen suggests novice users will be overwhelmed. How do you respond?
Sample Answer
Direct answer
A stakeholder is anyone with a say in or stake in the work, here the person pushing for the dense layout. I would neither refuse nor build it blindly. I would treat "a dense dashboard" as a solution the customer proposed and find the need behind it: which users, doing which decisions, how often. Then I would bring the stakeholder evidence rather than opinion, and propose a design that serves both the experts who want density and the novices who might drown, validated with a quick test before we commit.
Step by step
- Clarify who asked and why. Which customers made the request, and are they the same people as the novice users I am worried about? Power users who live in the tool all day want density. New users want a clear first answer.
- Find the decisions behind the density. Ask "what do you decide with this screen on a Monday morning?" If three questions drive most of the value, those need prominence; the other numbers can sit one click away.
- State my concern as a testable claim. "I expect first-time users to fail at finding X on the dense layout." That is something a test can show wrong, which keeps the conversation from becoming opinion against opinion.
- Prototype both and test with real tasks. Give representative novice users and experienced users the same two or three tasks on each version and watch where they stall. Agree in advance what would count as a pass (for example, most participants can find the key figure unaided).
- Offer alternatives that meet the request.
- A summary view by default with drill-down for detail (progressive disclosure: show the essentials first and reveal more on demand).
- A density toggle or saved "expert view" for power users.
- Role-based defaults, so each user lands on the metrics they own.
- Decide and record. Ship what the evidence supports, and write down what you would revisit.
Worked example
An operations dashboard has 24 tiles. Asking the requesters what they check first reveals that they open it to answer three questions: is anything late, is anything failing, what changed since yesterday. I would put those three at the top, keep the other 21 tiles in an expandable detail view, and test both layouts with a few new and a few experienced users. If experienced users complete tasks equally well on both and novices only succeed on the layered one, the layered version is the clear call.
What would change my mind
If testing shows the actual users are specialists who value seeing everything at once, I would ship the dense version, possibly with better grouping and labels. The aim is the customer's outcome, not winning the argument.
Pitfalls
Arguing from taste ("clutter is bad"), testing with colleagues instead of real users, and presenting the stakeholder's request as wrong rather than as a solution worth testing.
If you interviewed one of our customers, what core pains would you expect them to describe, and why do you believe those pains are relevant to the company's mission?
Sample Answer
Direct answer
Without having interviewed them, I can only offer hypotheses, not facts, and I would say so. I would build them from the company's product, its customers' jobs to be done (the outcomes customers hire the product to achieve) and public evidence, name the three or four pains I would expect, and then explain the link to the mission. The goal is to show I start from the customer's problem, then check my guesses in the first conversation.
How I would form the hypotheses before the interview
- Read what the company says it does and for whom (the mission and the target customer).
- Skim the places customers talk honestly: product reviews, support forums and community threads, job postings from customers' side, and competitors' reviews.
- Ask "what is the customer trying to get done, and what gets in the way?" That defines a pain point: a specific obstacle that makes a task slow, risky or frustrating.
Illustration (a made-up company: expense-management software for small businesses, mission "make spending simple and visible")
| Expected pain | What the customer would likely say | Why it matters to the mission |
|---|---|---|
| Chasing receipts | "I spend the end of every month asking people for photos of receipts." | Spending is not visible until someone reconstructs it |
| Slow reimbursement | "Staff wait weeks to be paid back." | Simplicity: the process creates friction for employees |
| Out-of-policy spend found late | "I only discover it when the card statement arrives." | Visibility comes too late to change behavior |
| Month-end close effort | "My accountant needs everything reformatted." | The work stays manual, so the product has not simplified it |
How I would tie each pain to the company's purpose
A pain is relevant to the mission if solving it moves the customer toward the outcome the company promises. If a pain is real but unrelated to the mission (say, complaints about a tax rule the company cannot influence), I would note it, but I would not prioritize it.
In the interview itself
- Ask for recent specific stories ("walk me through the last time you did month-end"), not opinions about features.
- Listen for severity (how bad), frequency (how often) and current workarounds, which show how much the pain is worth. Questions that draw each out: "What happens if the receipts are late?" (severity); "How many times did that come up last month?" (frequency); "What do you do today instead?" (workaround).
- Treat my hypotheses as something that could be wrong. A pain I did not expect is the most useful result.
Pitfalls
- Presenting guesses as findings.
- Listing solutions ("they want an app") instead of pains.
- Naming generic pains ("it is expensive") that would fit any company and show no research.
Unlock Full Question Bank
Get access to all 29 Customer and User Obsession interview questions and detailed answers.
Sign in to ContinueJoin thousands of developers preparing for their dream job.