Product research · Step by step

How to do product research: 7 steps that work in 2026

Product research takes one afternoon, not one month. Follow these seven steps in order and you finish with a product to test and a spreadsheet that shows why.

By Overview Data ·

Step 1: Choose one country before you choose a product

A product is only winning somewhere. Pick the single country you will sell in first. Prices, couriers, payment habits and competition change completely between Saudi Arabia, Algeria and France.

Write down: the country, the currency, the payment method (cash on delivery or card) and the couriers you can use. Everything after this step is checked against that country, not against "the internet".

Step 2: Build a shortlist of 20 candidates

Collect 20 products. Not 3, not 200. Twenty is enough to find two winners and small enough to finish today.

Pull candidates from three places:

  1. Ad library, filtered to your country. Search your category and note every product with a live ad. This is the best list because every item already has a seller paying to promote it.
  2. Best-sellers of competitor stores. Open the top shops in your niche and write down their featured products.
  3. Marketplace movers. Amazon Movers & Shakers and AliExpress top sellers for your category.

Put each candidate in a row: name, link to an ad, link to a store, country. Details on the sourcing methods are in how to find winning products.

Step 3: Check proof — how long have the ads run?

For each candidate, find the oldest live ad in your country and note how many days it has run.

  • 60+ days live: proven. Keep it.
  • 14 to 59 days: unproven. Keep it only if several sellers run it.
  • Under 14 days: cut it. You are looking at a test, not a winner.

Cut every candidate with no ad at all. "Nobody advertises it" does not mean "open opportunity". It means nobody found it profitable.

This single step removes about two thirds of any list.

Step 4: Count the sellers

Count the independent shops advertising the product in your country.

Sellers advertisingVerdict
1 to 3Too thin: could be one seller's lucky ad. Needs more proof
4 to 15Sweet spot: proven and not crowded
16 to 30Contested: enter only with a better offer or creative angle
30+Saturated: skip

Then open three of the competing pages and read them. Note the price, the offer (bundle, free delivery, guarantee) and what the ad shows in the first three seconds. That is the standard you have to beat. More on this in verify product saturation.

Step 5: Run the margin math

Take the lowest price competitors charge and calculate:

Landed cost  = product + shipping to your warehouse or courier
Break-even ad cost per order = Selling price − landed cost − delivery fee − packaging

The rule: selling price must be 3x landed cost or more. On cash-on-delivery, plan for 20 to 30 percent of orders never being collected and use 3.5x.

A $25 product with an $8 landed cost passes. A $25 product with a $14 landed cost fails, whatever the ads look like. Shopify's own worked example shows a $25 item keeping roughly $12.95 after fees, a 42 percent margin, and that is before ad spend (Shopify).

Also cut anything fragile, oversized, restricted (batteries, liquids, medical claims) or with a landed price under $5, where margin per order cannot cover ads.

Step 6: Confirm you can deliver it

For each survivor, answer yes or no to all four:

  • Do you have two suppliers, with quoted price and delivery time?
  • Can a courier deliver to your target cities?
  • Can you get a sample in your hands before you advertise?
  • Is the product legal to advertise on Meta in that country?

A "no" on any line kills the candidate. Supplier problems found after launch cost ten times more than supplier problems found today.

Step 7: Test the best 3, keep one

Your list is now down to 3 to 5 products. Rank them by margin and proof, take the top 3 and test each with the same small budget and the same offer structure.

Set the kill rule before you spend: if a product has not produced an order after your set budget, it is dead. Do not extend it. The winner is the product that gets orders at a cost per order below your break-even from step 5.

Then repeat the cycle every week. Ad libraries refresh weekly, and a product that passed today can be saturated next quarter.

The one-page spreadsheet

Make one sheet with these columns and fill one row per candidate:

ProductCountryOldest live ad (days)Sellers advertisingLowest competitor priceLanded costMultiple (price ÷ cost)Supplier readyDecision

Any row with days under 60, sellers over 30, multiple under 3 or no supplier reads Skip. Everything else is a test. You can run the first three columns for free with an ad library, or Search a product in Overview to see ad age and seller count for your country in one view.

Keep reading