Product Opportunity Desk
Product research this week: does this opportunity deserve another round of proof?
This week’s question is not whether there is another idea. It is whether one candidate deserves more time for a sample, page work, and a modest test budget. Put market research, demand validation, competitor review, margin and fulfillment screening, and weekly cadence into one sheet so the team can name the primary opportunity, backup, rejection reasons, and the one proof to add next week.
PetNest is assessing one rear-seat pet cover in the US market, not deciding whether the whole pet category is attractive. In 20 May 2026 mixed or negative reviews, one cleaning, fit, or dry-time problem must appear at least five times before a quick-release, washable, hair-resistant sample check. Research keeps proof, supply confirms sample and MOQ, and finance or ops checks cost and fulfillment. If any link is missing, there is no large order or traffic.
Working asset
Product opportunity evidence sheet
Pass standard
1 primary idea + 1 next action
Next lesson
Inventory planning
Opportunity status preview
Ready to test
Pet car-cleaning scenario with aligned search, review, and ad signals.
Evidence gap
Desk ergonomics has demand, but the persona and problem are still too broad.
Paused
Low-ticket trend accessory has attention, but margin and substitution risk are high.
Blocked
Certification-heavy, IP-risk, fragile, or return-heavy directions.
01 Correct the misread
If the meeting ends with keep looking, it produced no operating asset.
Heat is input. Evidence is judgment. A test path is action. A stop rule is risk control. You are not collecting inspiration; you are deciding what deserves the team’s next week.
Heat
Only proves people noticed.
Evidence
Explains why they would buy.
Test
Turns judgment into a small action.
Stop rule
Stops losses from scaling.
Do not rank candidates by which one feels most exciting. Check explanation cost, back-end complexity, conservative margin room, and expansion after validation: can the first screen explain it, can advertising demonstrate it, will supply and support slow the test, and can a win extend into a bundle or related use case? If one of those is unclear, add that proof first instead of presenting it as a higher-priority idea.
01A Count the review sample first
Classify the 20 reviews before asking a supplier to solve anything.
This step turns vague buyer complaints into a product hypothesis that can be assigned and reviewed. It does not prove that a pet car-seat cover will sell. It decides which two problems deserve a week of samples and cost checks.
PetNest starts with only 20 mixed or negative reviews from the US market in May 2026. It does not mix in five-star praise, other countries, or other vehicle types.Each review receives one primary tag: curled edges or poor fit 6, hair caught in seams 5, slow drying after washing 4, installation friction 3, and odor or durability 2. The arithmetic is 6 + 5 + 4 + 3 + 2 = 20, so the team can reconcile the total and spot a review that was counted twice.
In this case, five mentions is a working line for supplier review, not a platform rule for every store. Curled edges or fit is 6 >= 5 and hair caught in seams is 5 >= 5, so both enter the candidate queue. Slow drying is 4 < 5, installation is 3 < 5, and odor or durability is 2 < 5, so they stay in the evidence sheet instead of triggering a factory redesign. This line applies only to this small sample and this week.Recalculate it when the sample changes.
On Monday, the research owner saves review links, screenshots, dates, and tags. On Tuesday, the sourcing owner asks suppliers for sample proof of quick release, washability, and hair-resistant seams or structure. On Wednesday, the content owner turns the two evidenced problems into a before-and-after demonstration that can be shown in 15 seconds. On Thursday, finance or operations checks landed cost, packed weight, refund explanation, MOQ, and replenishment lead time. A sentence that a supplier can solve it is not evidence. Keep the sample version, quote date, weight or dimensions, and the owner who closes the candidate if it fails.
Friday also needs a bounded decision: if both frequent problems have reviewable samples and cost and fulfillment do not break the existing contribution-profit guardrail, move only to one small sample and one page-promise test. If the sample is missing, packed weight rises materially, MOQ locks cash, or the refund explanation becomes harder, return to evidence gap or pause. Twenty reviews prove a queue of problems worth validating. They do not prove revenue, advertising payback, or lasting product-market fit.
01B Review-coding calibration
Turn 20 review paraphrases into reviewable primary tags.
Agree on the coding rule before discussing frequency. Two reviewers may notice different concerns, but each row keeps one primary tag for this week’s next step.
The 20 lines below are de-identified classroom paraphrases built from this case’s counting logic. They do not expose an original reviewer, link, or platform, and they are not presented as a new market sample. Their purpose is to practice how a paraphrase, source location, date, and one primary tag can be reviewed.
Use a narrow coding rule: the primary tag names the problem that needs validation first. A reasonable concern from a second reviewer stays as a secondary note and is not counted again. The weekly review can then resolve the disagreement before confirming that 6 + 5 + 4 + 3 + 2 still equals 20.
| Row | Classroom paraphrase | Primary tag | Second reviewer’s first tag |
|---|---|---|---|
| After installation on a sedan rear seat, both edges curl upward. | Fit / edge | Agrees | |
| Even with the straps tight, the cover slides toward the center after turns. | Fit / edge | Agrees | |
| The curved side seats leave the edge unable to stay down. | Fit / edge | Agrees | |
| With a child seat in the rear, the attachment points do not feel stable. | Fit / edge | Agrees | |
| After one or two days, it wrinkles again and catches at the edge. | Fit / edge | Agrees | |
| Hair is hard to clean, but the first failure I notice is the curled edge. | Fit / edge | Hair / seams | |
| Dog hair keeps getting trapped in the center seam, and the vacuum cannot clear it. | Hair / seams | Agrees | |
| The seams are too deep; the more hair there is, the harder they are to clean. | Hair / seams | Agrees | |
| Hair remains in the seams after washing; the second reviewer first tagged it as slow drying. | Hair / seams | Slow dry after washing | |
| After every dog trip, I have to remove it and pick hair out of the seams. | Hair / seams | Agrees | |
| The surface wipes easily, but the seams are a hair trap. | Hair / seams | Agrees | |
| It is still damp the next day after machine washing and cannot be reused. | Slow dry after washing | Agrees | |
| It dries very slowly after washing on a rainy day. | Slow dry after washing | Agrees | |
| After hanging overnight, the back is still wet. | Slow dry after washing | Agrees | |
| The quick-dry promise did not hold, and it smells when damp. | Slow dry after washing | Odor / durability | |
| The setup diagram is unclear, so the first installation took a long time. | Installation friction | Agrees | |
| The buckle locations are hard to find, and I had to restart after installing it. | Installation friction | Agrees | |
| It is awkward for one person to install in a car because there are too many steps. | Installation friction | Agrees | |
| It smells on unboxing, and I worry the odor will stay in the car. | Odor / durability | Agrees | |
| After a few weeks it smells, and the edges are starting to become brittle. | Odor / durability | Agrees |
These 20 lines are classroom calibration material, not original market evidence. Even fully consistent coding does not prove demand, revenue, profit, or product approval.
02 Product opportunity decision practice
Decide the candidate status before deciding whether to test.
This keeps the lesson from staying at principle level. Read What / Why / How / Practice, then choose a real candidate and this-week action to see whether it can be written back to the evidence sheet.
What
An evidence sheet is not a product list
It records why a candidate deserves a test, what evidence is missing, who will fill the gap, and when to stop. Without those fields, even a hot product is only inspiration.
Why
A store sells an explainable opportunity, not a marketplace copy
Marketplace volume, social heat, and review count only prove attention. A direct store must also prove cold traffic can understand it, margin can carry it, and fulfillment can keep the promise.
How
Route the status before assigning a test
Every candidate must enter one of four states: ready to test, evidence gap, paused, or blocked. If the status is unclear, do not move into samples, page work, or ad budget.
Practice
Turn this week’s candidate into one evidence line
A valid line includes audience, problem, first evidence, status, next-week action, and stop rule. If you cannot write it, fill the evidence gap first.
The exercise below is not asking you to pick the most attractive-looking product. It checks whether you can match a status to its evidence. Read each candidate situation and its first proof, then choose this week’s action. The feedback explains why margin, review, supply, or compliance evidence comes next.
Use the PetNest rule as the reference: passing review classification does not authorize inventory, a supplier sample does not prove fulfillment, and a clear page angle does not make contribution profit positive. Fill one gap that changes the decision at a time so a low-cost judgment does not become a full page, shoot, and advertising spend.
First evidence
Car-cleaning kit went viral
First collect conservative price, landed cost, expected CPC, package weight, refund/reship rate, and minimum contribution margin.
After the exercise, write the candidate name, current state, first gap, owner, due date, and stop line. If any field cannot be written, do not interpret the result as test-ready. The correct next action is to keep filling evidence.
03 Main artifact
Break I think it can sell into 12 evidence fields.
This product opportunity evidence sheet is the core artifact. Click a field to see the question, evidence, fail signal, and owner.
Active field
Target audience
Question
Who repeatedly has this problem, and in which moment does it hurt most?
Evidence
Search phrases, review language, social comments, competitor FAQs, and real use cases.
Fail signal
If everyone could buy it is the answer, the audience is not defined.
Owner
Merchandising owner
04 Market Problem Map
Start with audience and problem. The SKU appears last.
Fall in love with a SKU first and you will reverse-engineer excuses. Start with audience and problem so ads, pages, support, and fulfillment can connect.
Audience
Pet owners
Problem
Travel, car rides, cleaning, and storage create repeated friction.
Solution type
Tools, kits, cleaning consumables, travel organizers, and scenario-specific promises.
Candidate
Test car-cleaning or travel organization first; do not expand into a full pet catalog yet.
05 Signal Cross-Check
The five signal groups are not a source list. They cross-check each other.
A single source lets weak ideas survive. For each signal, write support, counter-signal, and next move instead of just saving screenshots.
Search signal
Weekly source list
Start the weekly review by checking source, responsible team, and data date. A judgment without a source list stays in evidence gap before page work, samples, or ad budget.
Reviews / support tags / refund reasons
Responsible team: Support + content team
Cadence: Update weekly and keep the data date.
At least 20 classifiable reviews, tickets, or refund tags grouped by cleaning, fit, smell, durability, setup, or return reason.
Search terms / site search / Search Console queries
Responsible team: SEO team
Cadence: Add 10 related terms weekly and tag learning, comparison, buying, or problem-solving intent.
Split object, constraint, context, and result, and capture site-search queries where buyers could not find a match, such as desk cable / no drill / renter desk / hide power strip.
Ad libraries / short video / creative comments
Responsible team: Growth team
Cadence: Record five long-running ads or reusable demos. Do not treat one spike as entry permission.
Capture first line, demo action, landing promise, price band, and buying questions in creative comments.
Shopify products / orders / Feed fields / supplier changes
Responsible team: Operations + merchandising + supply team
Cadence: Review before each small test; tool-source checks were reviewed through 2026-06-27.
Record Shopify product and order evidence, title attributes, variant, GTIN/MPN, product_type, custom_label, availability, sample consistency, MOQ, and reorder lead time.
If a candidate still depends on offline service or fulfillment, an online signal is not a complete viability proof. An exploratory interview study of online-to-offline ecommerce found that interviewed managers linked online and offline resources as complementary, treated offline operations as a continuing source of value, and associated O2O with broader geographic reach. It covers managers across Chinese service and commerce industries and cities; its limited sample and convenience/snowball recruitment restrict transfer, and it does not prove revenue, profit, cost, or retention lift. In the evidence table, record the online demand signal, the offline resource that must deliver the promise, an owner, and a date. If one is missing, keep the item in evidence gap or pause and collect local merchant and customer data.
Evidence quality check
Sample size and dates are not paperwork. They stop the team from treating one review, one ad, or one short-video spike as a product opportunity.
Review sample
At least 20 classifiable reviews with source and date.
One or two quotable lines do not prove repeated pain.
Search terms
At least 10 related terms tagged by intent and scenario constraint.
One hot term may be news, entertainment, or broad interest.
Competitors and ad continuity
At least five serious competitors or five continuity signs from ads.
Only reading one leading brand can misread price band and openings.
Supply and math date
Supplier replies, sample proof, and conservative math version need dates.
Stable supply or profit works is not reviewable.
06 Opportunity signal mining board
Mine a product opportunity from reviews, search terms, and ad language.
Do not stop at I found an idea. The operating move is to translate buyer words, usage constraints, and repeated competitor ad language into a candidate product, page promise, and this-week evidence action. Pick a signal path below.
Before switching the signal paths below, separate raw signal from conclusion. Buyer wording in reviews, constraints in search terms, and repeated ad language are inputs only. A product hypothesis must name the audience, use case, demonstrable improvement, and who will fill which evidence by when.
For example, six fit complaints and five hair-in-seam complaints can lead to a quick-release, washable, hair-resistant hypothesis for the pet car-seat cover. They cannot lead directly to a large order. Record whether competitors already solve it, whether supplier samples are consistent, and whether packed weight and return cost remain inside this series financial guardrail.
Active mining path
Mine opportunity from reviews
Write back: Review opportunity: pet car cleaning | Buyer words: hair is hard to clean / slow to dry | Hypothesis: quick-release washable structure | This week: review clustering + supplier structure check.
Do not treat the opportunity hypothesis on the right as approval. Copy the raw signal, problem tags, supplier evidence request, and this-week owner into the evidence sheet. If sample, cost, or page-promise proof is still missing next week, keep the status at evidence gap instead of opening ads.
07 Question quality check
Good product opportunities often show up as repeated buyer questions first.
A University of Toronto paper on agentic purchasing frames shopping as a path of asking before recommending. In product research, do not only ask whether a trend is hot; ask what buyers keep asking, whether product data can answer it, and whether the page has proof.
Use question quality to screen opportunities
If an opportunity cannot answer the questions buyers repeatedly compare, worry about, or use to choose, heat alone does not make it test-ready.
Buyers repeatedly compare specs, size, material, or use cases.
Evidence to add: Attributes, usage scenes, comparison visuals, and FAQs need to answer directly.
Pass standard: The page answers the main questions in the first screen or the next two blocks.
Buyers first ask whether it solves one concrete friction.
Evidence to add: Before/after proof, demos, reviews, and risk reversal must support that problem.
Pass standard: Each key claim has page evidence and support language behind it.
Buyers need guidance on version, bundle, or configuration.
Evidence to add: Buying Q&A, size/config tables, and recommendation rules need to be prepared.
Pass standard: Ads, page copy, and support give the same choice logic.
08 Competitor opening
Competitor analysis is not copying pages. It finds where you can enter.
Review at least 5-10 serious competitors. You are not looking for their theme; you are looking for unclear audience, scenario, bundle, or trust support.
Price band
Question: Can you avoid pure low-price competition, or only be cheaper?
Output: Record transaction range, bundles, free-shipping threshold, and discount frequency.
First-screen promise
Question: Is every competitor saying the same thing?
Output: Find one more specific line that explains scenario, outcome, or risk reversal.
Offer structure
Question: Is there room for kits, upgrades, accessories, or bundles?
Output: Do not copy SKUs. Decide if the buyer needs a more complete solution.
Trust support
Question: Are shipping, returns, reviews, FAQs, and support clear?
Output: Turn competitor ambiguity into your page evidence and support language.
09 Feasibility Gate
Do not use ads to scale an opportunity the back end cannot support.
Margin, fulfillment, returns, compliance, supply, and message cost need screening during research. Failure is not only zero demand; sometimes orders are unhealthy.
Cash flow is the money that actually enters and leaves the store over time: orders and payment payouts come in, while purchasing, ads, shipping, refunds, and reships go out. It is not the same as gross margin. A candidate can appear profitable while MOQ, replenishment timing, or upfront ad spend ties up cash before the test has a conclusion.
The order is therefore: first confirm that the product and page can carry the buyer problem, then put profit, fulfillment, and cash tie-up into the downside case together. If any part misses the line, return to evidence gap or pause instead of using a larger budget to force a verdict.
Read the common fields in plain language first. MOQ (minimum order quantity) on a supplier quote tells you how much stock must be committed, not whether quality or timing is reliable. CPC (cost per click), CVR (conversion rate), and CPA (cost to acquire an order) appear in ad-platform and analytics reports to estimate traffic cost, not to prove durable profit. GTIN/MPN are product identifiers used for feed and catalog matching, not proof that a product has passed a channel review.
Keep digital adoption separate from realized profit. A study of Vietnamese online-retailing firms used a cross-sectional questionnaire and structural equation modeling. Of 432 completed questionnaires, filtering left 325 valid firm responses; the survey ran for 7 weeks, and the paper does not report calendar dates. It reports a positive modeled relationship between online-retailing adoption intention and perceived business performance. That is not audited sales or profit, and it is not causal evidence. Write compatibility, observable benefits, top-management support, and technical or legal conditions as local hypotheses to test. Next to product-data readiness, record the expected improvement, reviewer, and evidence date. If one is missing, return to evidence gap or pause and collect local order, fulfillment, and profit data instead of treating the paper's relationship as proof of profitability.
Margin
Stop: No testing room after ads, payment, shipping, refunds, and reships.
Go when: Conservative margin can absorb first-round failure cost.
Fulfillment
Stop: Heavy, bulky, fragile, or remote shipping cost is uncontrolled.
Go when: Delivery time, exception handling, and return path are clear.
Compliance
Stop: Certification, IP, children, electronics, battery, or medical boundary is unclear.
Go when: Sellable boundary, page wording, and supplier evidence can be kept.
Message
Stop: The first screen cannot explain why buyers buy, and creatives rely on adjectives.
Go when: Buyers can understand problem, result, and reason to buy within 10 seconds.
Product data readiness before test
During product research, check whether product data can be reused by Merchant Center, Meta Catalog, collection filters, and inventory management.
title attributes
Can the title carry material, use case, fit, size, or bundle facts?
variant option
Do color, size, bundle, and fit options affect inventory and returns?
GTIN / MPN / SKU
Are GTIN/MPN status, SKU naming, and later Feed attribution clear?
product_type / custom_label
Can Merchant Center, Meta Catalog, collection filters, and budget labels reuse it?
availability / image proof
Can availability, price, sale_price, and main-image proof support the test?
Supplier evidence row
Supply stability is not a sentence. It needs to be reviewable and reusable after a winning test.
sample photo / video
Are sample photos, videos, and bulk consistency promises saved?
MOQ / lead time / reorder lead time
Will MOQ, first lead time, and reorder lead time pressure cash flow?
defect policy / packaging weight
Are defect handling, package weight, damage, and reship rules clear?
backup supplier / QC checkpoint
Is there backup supply or a clear reopen condition, plus QC checkpoints?
Conservative test math
Run a downside case before testing. Do not look only at gross margin or marketplace volume.
landed cost / shipping / payment fee
Count landed cost, shipping, payment fee, and discount-adjusted unit cost together.
refund / reship reserve
Reserve a downside case for refunds, reships, damage, and support compensation.
expected CPC / CVR / CPA
Use conservative CPC, CVR, and CPA to decide whether test room remains.
contribution profit / cash tied in MOQ
If minimum contribution profit and cash tied in MOQ fail, do not enter ads or inventory.
10 Opportunity stop-rule router
Heat does not mean an idea is test-ready.
This step answers why an idea should not be tested yet. It is not about being conservative; it checks hard risks before page work, creatives, samples, and ads.
Opportunity stop-rule router: heat does not mean test-ready
This step answers why an idea should not be tested yet. It is not about being conservative; it checks hard risks before page work, creatives, samples, and ads. Pick a scenario below to align symptom, hidden risk, first proof, repair action, and the line to write back into the evidence sheet.
High heat, thin margin
Surface symptom
TikTok or ranking signals look strong, but competitor prices are low, clicks are costly, and shipping plus after-sales costs are not fully counted.
Hidden risk
Orders are not profit. Without margin room, even a small test can increase cash and support pressure.
First proof
First write conservative price, landed cost, payment fee, expected CPC, refund/reship rate, and minimum contribution margin.
Repair action
If margin is too thin, change price band, bundle, supply cost, or market. Do not replace math with let us test it.
Route
Pause until margin is repaired
Do not
Do not build a full page, place a large sample order, or launch many ad sets.
Write back: Write back: margin below test line; resume only when contribution margin reaches the test line or a higher-price/lower-cost route is found.
11 Weekly cadence
Continuous product research is not one research sprint. It is a weekly operating rhythm.
Each week should produce candidates, validation, rejections, and review notes. Next week starts from evidence, not another brainstorm.
Monday
Collect candidate directions
Add 5-10 opportunities
Tuesday
Cross-check five signal groups
Mark support and counter-evidence
Wednesday
Review margin, fulfillment, compliance
Flag back-end risks
Thursday
Prioritize tests
Choose 1-2 testable opportunities
Friday
Review and reject
Document continue, pause, or reject reasons
12 Weekly opportunity review card
In the Friday review, do not stop at this product looks decent.
Choose the closest state for this week’s candidate, then write the stop rule, next action, and next-lesson route into your copyable notes. Next week, the team knows whether to fill evidence, check profit, inspect inventory, or run a page test.
Choose this week’s state
Click a card and the right side becomes a review note you can use in the weekly meeting.
Weekly opportunity review card
Fill the evidence gap
Current evidence
There is heat, review, or search signal, but the audience is broad and the page angle is not tied to one concrete pain.
Evidence quality
Fewer than 20 review samples, fewer than 10 search terms, and competitor openings are not supported by at least five serious competitors.
Product data readiness
title attributes, variant option, GTIN/MPN, product_type, custom_label, and image proof cannot be written yet.
Supplier evidence
Supplier evidence is missing sample consistency, MOQ, lead time, packaging weight, and backup supplier proof.
Conservative test math
No downside case yet for landed cost, shipping, refund/reship reserve, payment fee, and expected CPC/CVR/CPA.
Stop condition
Do not invest in page, creative, or samples yet. If next week still cannot define the problem, downgrade it in the pipeline.
Next-week action
Add three review pains, two search terms, and one competitor gap, then return to product validation scoring.
Write back to review note
Review state: fill evidence first; no page or ad work this week, only review pains, search terms, and competitor gaps.
Tip: when you click “Copy lesson notes” below, this review card’s stop condition, next action, next route, and review note line are included automatically.
Classroom review record
Save this classroom review without turning it into a product approval.
Saved or exported from this browser only.
The record keeps the current review card, triage state, opportunity-practice choice, 20-line classroom coding, and four evidence checks. It does not collect or retain an account, order, customer, real review link, supplier contact, quote, actual budget, or platform data.
Review coding
0/20 match; 0 need review
Evidence check
0/4 source-and-date checks marked
Opportunity practice
No this-week action is selected yet.
Keep filling evidence or correcting coding. Prepare for a human weekly review only after classroom coding, source-date checks, and the current case action all pass.
13 Opportunity Triage Board
Every candidate needs a state. They cannot all be keep looking.
The opportunity state decides the next action. Click a state to see entry rule, common misread, next move, and the field to write into your copyable notes.
Testable
Entry rule
Signals align, back-end risk is controlled, and a low-cost two-week test is possible.
Common misread
Treating testable as scalable and expanding SKUs or inventory immediately.
Next move
Use one market, one message, one creative set, and one stop rule.
Notes field
Test hypothesis
Quick Check
A product is hot on TikTok and sells on Amazon, but margin is thin, return reasons repeat, and competitor pages compete on low price. What should you do now?
Use this lesson’s order before answering: 20 classifiable reviews, two primary problems at five mentions, supplier sample proof, and cost and fulfillment review come before the next step. One hot term or one competitor ad is not substitute evidence.
Whatever answer you choose, write the result as a state rather than a feeling: evidence gap, paused, blocked, or ready for a small test. Then assign an owner and due date. Do not change the state without new reviewable evidence.
14 Copyable lesson notes
Turn this lesson into copyable product research notes.
When merchandising, content, media, supply chain, and operators read the output, do not stop at this product looks good. Copy evidence, counterevidence, cash-flow risk, test action, owner, and stop rule.
Copyable lesson notes
Complete these notes before inventory, pricing, visual content, conversion, or feed operations. If you cannot fill them, the opportunity is not ready.
Where should I go next?
Opportunity exists, but inventory is unclear
Opportunity passes, but pricing is unclear
Opportunity needs content validation
Page expression and conversion are weak
Product data readiness is weak