01What is a winning product?
"Winning product" gets used like a lottery ticket: one magic item that sells itself. In practice, a winning product is simply a product that real people want, that you can show convincingly in an ad, that leaves enough margin to pay for customer acquisition, and that you can deliver reliably. When all of those are true at the same time, you have something worth testing. When one is missing, the product usually loses money no matter how viral it looks.
That definition matters because it shifts the question. Instead of asking "what's trending?", you ask "which of these products can I sell profitably to a specific person, with the budget and suppliers I actually have?" A product that works for one store may fail for another because the price point, market, creative skills or supplier are different.
Product research doesn't find guaranteed winners. It improves the odds and makes your tests cheaper and faster. The goal is a repeatable process that keeps producing products worth testing.
Winning for whom, and at what price?
Always attach three things to a product idea: the customer (who buys it), the problem or desire (why they buy it), and the price (what they'll pay). "Posture corrector" is a product. "A posture brace for desk workers whose neck aches by the afternoon, sold at around $40" is something you can research, price and advertise.
02The 10-filter product research framework
Every idea goes through the same ten filters. The first few kill most weak ideas in minutes, so you only spend real time on products that deserve it. Mark each filter pass, maybe or fail. A single hard fail on demand, economics or delivery usually ends the idea.
- MarketFILTER 01Clear problem or desireCan you describe why someone buys it in six words or fewer?
- MarketFILTER 02Demand evidenceAre people buying it now, not just watching videos of it?
- MarketFILTER 03Audience sizeIs the group of likely buyers big enough to advertise to?
- MarketFILTER 04Competition & saturationIs there room for a new angle, offer or audience?
- MoneyFILTER 05Margin & economicsDoes the price leave enough room to pay for a customer?
- MoneyFILTER 06Perceived valueDoes it look worth much more than it costs you?
- CreativeFILTER 07DemonstrabilityCan a silent video show the benefit in a few seconds?
- CreativeFILTER 08Angle depthCan you write at least five different ad angles for it?
- DeliveryFILTER 09Supplier & shippingCan you get consistent quality to customers in a reasonable time?
- DeliveryFILTER 10Risk & returnsLow breakage, simple sizing, no restricted claims or IP issues?
The filters fall into four groups: market (does anyone want it), money (can it pay for itself), creative (can you sell it in an ad) and delivery (can you fulfil it without drama). Most beginners only check the creative group, because a good video is exciting. The market and money groups are where most products actually fail.
How the filters narrow your list
A typical research sprint starts wide and ends with a short list you can test properly. The exact numbers depend on you; the shape is what matters.
Keep every rejected idea in your sheet with the filter it failed. Weeks later, a better supplier, a new angle or a bundle can flip a "fail" into a "test".
03Where to research product ideas
Good ideas come from places where you can see real buying behavior or real advertising spend. Spend a focused 60–90 minutes collecting ideas from several sources before judging any of them. Capture the product, the link, the apparent price, and the angle being used.
| Source | What to look for | Why it's useful |
|---|---|---|
| Meta Ad Library | Search problem keywords ("back pain", "pet hair", "sleep"). Open ads that have been running for a long time; note how many advertisers sell the same product. | Advertisers rarely keep paying for losing ads for weeks. Longevity is a demand signal. |
| TikTok Creative Center | Top ads and trending products in your target country and category. | Shows current formats and angles, not just products. |
| TikTok & Instagram search | Phrases like "TikTok made me buy it", "amazon finds", niche problems; read comments for "where do I buy this?" | Comments reveal intent and objections. |
| Marketplace bestseller lists | Amazon Movers & Shakers, category bestsellers, AliExpress top sellers. | Evidence that people are paying, plus price references. |
| Google Trends | 5-year trend for the product and its problem keywords. | Separates steady demand from a one-week spike, and reveals seasonality. |
| Reviews and forums | 1–3 star reviews of existing products; Reddit and niche communities. | Shows unmet needs and what to fix or promise. |
| Your own world | Problems you, your family or your clients pay to solve. | You understand the customer, which makes better ads. |
Research ideas, not just items
Look for problems that keep appearing across different products: bad posture, pet hair, poor sleep, messy cables, sore feet. A recurring problem points to a niche, and a niche gives you several products, bundles and angles instead of one lucky item.
04How to validate demand
Views are not demand. A video with millions of views tells you people stopped scrolling; it tells you nothing about whether they paid, how much, or whether the seller made a profit. Before a product goes further, look for at least two of these three signals:
- Longevity. The same product (or close variants) has been advertised for 30+ days by at least one store. In Meta Ad Library, check the "started running" date and whether the advertiser keeps producing new creatives for it.
- Multiplicity. Three or more independent stores sell it with different creatives and pages. One store might be burning money; several stores investing is a stronger sign.
- Buying evidence. Marketplace listings still collecting fresh reviews this month, rising or stable Google Trends over a long window, or an active category bestseller ranking.
Check the trend shape
In Google Trends, look at five years, not ninety days. A flat or rising line suggests evergreen demand. A spike that collapses suggests a fad. A line that peaks every November or December is a seasonal product: fine for a Q4 push, risky as the only product in a store.
A single viral post with no ongoing advertisers and no marketplace sales is not validation. Treat it as a lead to investigate, then run the rest of the filters.
05How to analyze competition
Competition is not automatically bad. Several sellers usually means there's demand. The real question is whether you can offer a different angle, a better offer or a better experience than what's already out there.
Visit the three to five strongest stores selling the product and record:
- Price and offer: single price, bundles, free shipping threshold, guarantee.
- Angles: who their ads speak to and what problem they lead with.
- Page quality: images, reviews, shipping clarity, speed on mobile.
- Creative volume: how many different ads they run and how fresh they are.
- Customer complaints: comments and reviews mentioning delivery, quality or fit.
Saturated or still open?
| Signal | Still open | Likely saturated |
|---|---|---|
| Angles | Most sellers use the same one or two angles; obvious audiences are untouched | Every angle and audience you can think of is already being run |
| Prices | Stable across sellers | Racing down; frequent deep discounts |
| Creatives | Few sellers produce original content | Everyone reuses the same video |
| Retail | Not easy to find in local shops | Cheap in supermarkets or on Amazon |
If a product looks crowded, list five new audiences or problems it could serve before you drop it. A neck massager sold for "relaxation" competes with everything; sold "for nurses after a 12-hour shift", it speaks to one specific person.
06How to check product economics
This is the filter that saves the most money. If you run paid ads, every sale has an acquisition cost. You need to know the maximum you can pay for a customer before you test, not after.
Landed cost = Product cost + Shipping to customer + Packaging
Break-even CPA = Price − Landed cost − Payment fees − Other per-order costs
Example: $39.95 − $10.00 − $1.46 (2.9% + $0.30) − $0.50 = $27.99 you can spend to acquire one customer before losing money on the first order.
Break-even ROAS = Price ÷ Break-even CPA
Example: $39.95 ÷ $27.99 ≈ 1.43. Below that, ads lose money on the first order. Aim higher to cover refunds, overheads and profit.
Fee rates vary by payment provider, plan and country, so use your own numbers. As a starting rule, many operators look for a selling price of roughly three times landed cost for products sold with paid traffic, and a break-even CPA they believe they can realistically beat.
Use bundles to improve the math
A 2-pack or 3-pack at a per-unit discount often earns much more margin per order than a single unit, while the acquisition cost stays similar. If a product's single-unit margin is thin, check whether a bundle, refill or accessory changes the picture before rejecting it.
Check the market price range first. If competitors sell for less than about 2.5× your landed cost, either find a better supplier, build a stronger offer, or move on.
07How to evaluate creative and ad potential
On Meta and TikTok today, broad targeting means your creative largely decides who sees the ad and who buys. A product with great economics but nothing visual to show is hard to sell with short-form video.
Score each product on these questions:
- Is the benefit visible? Before/after, transformation, satisfying use, or a clear problem being solved.
- Does it work with sound off? Many people watch muted. The first frames should make sense silently.
- How many angles can you write? Aim for at least five: pain, desire, identity ("for nurses"), objection ("invisible under a shirt"), comparison.
- Can you get fresh footage? Can you film it yourself, get UGC creators, or get original footage from the supplier? Reusing the same video as everyone else rarely works for long.
- Is it easy to explain in one line? If you need a paragraph, consider an advertorial or longer landing page.
Write three hooks for the first three seconds of an ad. If you struggle to write even three, the product may be hard to advertise at scale.
08How to evaluate suppliers
Customers never meet your supplier, but they feel every late parcel and every broken item. Delivery speed and extra costs are among the most common reasons shoppers abandon checkout in Baymard Institute's research, and after the sale, supplier quality drives refunds and reviews.
Judge suppliers on five pillars, not unit price alone:
| Pillar | What to check |
|---|---|
| Price | Unit price plus shipping to your market, at your expected volume |
| Speed | Processing time plus shipping time; tracking that updates |
| Quality | Sample condition, packaging, consistency across orders |
| Communication | Reply speed, clear answers, willingness to send photos or video |
| Reliability | Store age and ratings, stock levels, history with disputes |
- Shortlist 3–5 suppliers for the same product (marketplace sellers, a sourcing agent, or an app like DSers or CJdropshipping).
- Send the same first message to each: processing time, shipping line and time to your market, and whether they can ship without price invoices. Note how fast and clearly they reply.
- Order samples from the top two or three using the shipping method your customers will get, and record real delivery days.
- Pick a primary and a backup, and write the terms down.
Show realistic delivery times on your product page and in shipping rate names. Customers who know the delivery window up front are far less likely to open "where is my order?" tickets or chargebacks.
09Common product research mistakes
| Mistake | Why it hurts | Do this instead |
|---|---|---|
| Chasing views | Views don't show purchases or profit | Look for ad longevity and multiple advertisers |
| Checking margin after launch | You discover CPA is higher than margin after spending | Calculate break-even CPA before testing |
| Testing one product for weeks | Slow learning and emotional attachment | Test 2–3 products in parallel with written kill rules |
| Copying the same viral video | Audience fatigue and possible copyright problems | Film or source original creatives with new angles |
| Ignoring delivery and returns | Refunds and bad reviews eat margin | Sample first; prefer durable, one-size products early |
| Trusting "winning product" lists | Everyone else received the same list | Use lists only as inputs to your own filters |
| Making claims you can't support | Ad rejections, refunds, legal risk | Describe real benefits; avoid medical or guaranteed-result claims |
10Product validation checklist
Run every product through this list before you spend on ads. Tick items as you go (your ticks stay on this device only).
11FAQ
How long does product research take?
A focused first sprint usually takes one to two weeks: a few days to collect ideas, a few days to filter and validate, and a few days to price, score and prepare tests. After that, two short research sessions per week keeps your pipeline full.
Is a saturated product always a bad choice?
No. Many sellers is a sign of demand. A product is only truly saturated when the audience has seen the same angles repeatedly and prices are racing down. A new angle, audience, bundle or better offer can still work.
What margin do I need to run paid ads?
Work out your break-even cost per acquisition: selling price minus landed cost, payment fees and other per-order costs. If that number is small, you have little room to pay for customers. Many operators aim for a selling price around three times landed cost for paid-traffic products, but calculate your own numbers.
How many products should I test at once?
Testing two or three products in parallel with fixed budgets and written kill rules is a common approach. It speeds up learning and stops you from getting emotionally attached to one product.
Can a product be guaranteed to win?
No. Research reduces risk; it doesn't remove it. Results depend on the offer, creatives, store, budget, timing and competition. The goal of a framework is to make better bets and learn faster.
Do I need paid tools for product research?
No. Meta Ad Library, TikTok Creative Center, Google Trends and marketplace bestseller lists are free and cover most of what you need. Paid spy tools can save time, but they don't replace your own validation.
Want the complete product research system?
This guide gives you the framework. The DaviEcom How To Find Winning Products Blueprint goes deeper into the full research workflow: the 10-point product scorecard with decision bands, the max-CPA calculation, worked product examples, a 14-day research sprint and a testing checklist, all in one structured PDF.
Preview the curriculum before you buy. Blueprints are manually delivered in your WhatsApp, Instagram or Telegram chat after payment verification.
KEEP LEARNINGMore from the DaviEcom Blueprint Library
Product research is step one. The DaviEcom Course covers the rest of the system in focused blueprints, from store building to ads.