ChatGPT Ads for Comparison Questions: When Shoppers Ask Which to Buy
Many shopping questions in ChatGPT are comparisons. Here is how to answer "which should I buy" with a specific product, a fair page and copy you can stand behind.
A large share of shopping conversations are not "what is this product." They are "which one should I get." Someone asks which sheets sleep cooler, which dog harness fits a puller, which espresso machine is simpler for a small kitchen, or whether a foam or hybrid mattress makes more sense for a side sleeper. The person is already choosing. They want a way to decide.
That moment is a strong fit for a chat card, if the card respects how the question works. A comparison question is a request for criteria, tradeoffs and a fit for their situation. It is a weak fit for a slogan, a superlative, or a card that names a rival brand. This guide covers how to find comparison questions worth advertising against, how to write context hints and chat cards for them, and how to build a landing page that finishes the comparison honestly.
What a comparison question is asking for
Read a comparison question as a short brief. It usually contains three things:
- A job to be done. Cooler sleep, a quieter commute, a stain that lifts, a gift that will not be returned.
- Constraints. Budget, size, material, skin type, pet size, apartment storage, a preference they already have.
- Options they are weighing. Two of your variants, two materials, or a category they do not fully understand yet.
Your ad does not need to settle every option on the internet. It needs to be a relevant next step for one of those constraints. "Percale Sheets for Hot Sleepers" is a direct answer to a cooling question. "Our Best Sheets" is not, and "best" is a claim you would have to prove. Our guide to ChatGPT ad headlines covers how to stay specific inside a 3 to 50 character title.
People also ask comparison questions in different shapes:
| Shape | Example | What the card should do |
|---|---|---|
| Material vs material | "Linen or cotton sheets for summer?" | Name the material you sell and who it suits |
| Use case | "Espresso machine for a small apartment?" | Name the constraint (size, noise, warmup) |
| Feature tradeoff | "Lightweight stroller or one with a bigger basket?" | State the tradeoff your product actually makes |
| Fit and body | "Running shoes for wide feet?" | Name the fit, not a ranking |
| Budget band | "Good chef's knife under a set budget?" | State price or a range you can honor |
| Your line vs your line | "Which of your candles is less smoky?" | Point to the variant that matches |
The buyer question finder is a useful way to draft this list from your category. Pair it with support tickets, review language and the questions you already answer on product pages. Our guide to the questions buyers ask before they purchase walks through the research.
Do not turn the ad into a competitor attack
Comparison questions tempt brands to write "better than Brand X" or to put a rival's name in the headline so the card feels relevant. That is a poor trade.
- You usually cannot see the exact prompt. OpenAI's policies prohibit ads that imply you know what the person said. "Since you are comparing us to..." is the kind of line that fails that rule. Our ChatGPT ad policy guide covers the separation between the answer and the sponsored card.
- Comparative claims need proof. "Lasts longer than the leading brand" is a factual claim. If you do not have a fair test, do not say it. Our ad claims compliance guide explains substantiation.
- Names and logos create legal and policy risk. Using another company's name, packaging or trademark in the card or image can get the ad rejected and can create trademark problems. The ChatGPT ad images guide flags competitor products in the frame as a rejection risk.
- Superlatives are hard to support. "The best," "number one" and "most popular" need evidence that matches the exact claim. A review quote you can attribute is safer than a ranking you invented.
A useful card answers with your product's facts: material, size, who it is for, what is included, the offer. The shopper can compare those facts to whatever else they are reading. You do not have to declare a winner.
Context hints for decision moments
Context hints describe the shopper's situation in plain language. For comparison traffic, describe the decision, not a brand war.
Stronger hints:
- "Person deciding between linen and cotton sheets for hot sleep"
- "Apartment dweller comparing compact espresso machines"
- "Runner with wide feet looking for a daily training shoe"
- "Dog owner comparing harnesses for a dog that pulls on walks"
- "Home cook choosing a first chef's knife and wanting to understand steel types"
Weaker hints:
- "Person searching for the best sheets"
- "Someone who should switch from a competitor"
- "User comparing our brand to a named rival"
The weak versions are broad, hostile, or they pretend the ad is inside the conversation. Write one hint per real decision, then point that hint at the product or page that fits it. Our context hints guide and the context hints generator help you draft and then edit them.
Example chat cards
Chat cards are short: a title of 3 to 50 characters, a body of about 100 characters, an image and a link. The examples below are for hypothetical brands. Swap in your real materials, prices and policies. Check length with the ad character counter.
Sheets
- Title: Percale Sheets for Hot Sleepers
- Body: Crisp cotton percale, not a brushed finish. Queen and king. Free shipping over the threshold shown at checkout.
Espresso
- Title: Compact Machine, 9-Inch Wide
- Body: Fits a small counter. 25-second heat-up. Includes a milk pitcher. 30-day returns.
Shoes
- Title: Daily Trainers in a Wide Fit
- Body: Same last in wide sizes. Neutral cushioning for road miles. Free exchanges on size.
Harness
- Title: Front-Clip Harness for Pullers
- Body: Front leash clip for dogs that lunge. Three sizes by chest girth. Reflective trim.
Knife
- Title: 8-Inch Chef's Knife, German Steel
- Body: Full tang, 8-inch blade. Includes a sheath. Sharpening guide on the product page.
Notice what these cards skip. They do not say "better than," they do not name another brand, and they do not claim a clinical or guaranteed outcome. They give the shopper a fact they can use in the comparison they are already making. More patterns are in our chat card copy examples and body copy guide.
Landing pages that finish the comparison
The click is the start of the comparison, not the end. If the card says "percale for hot sleepers" and the page opens with a brand film, you dropped the thread. A comparison-ready page usually includes:
- A first screen that restates the decision. "Percale for hot sleepers" should be visible without scrolling.
- The criteria, in plain language. Weave, weight, feel, who tends to prefer it, and who might want a different fabric you also sell.
- Your own lineup, compared fairly. A table of your variants (percale, sateen, linen) with the same rows for each is more useful than a vague "why us" section. Keep the rows factual: material, weight, care, price.
- The offer and the risk. Shipping, exchanges and returns belong next to the choice, because comparison shoppers are still unsure.
- Proof you can source. Reviews that mention the exact criteria ("sleeps cooler," "true to the wide fit") help more than a star average with no words.
If you sell several products that answer different questions, send each card to the matching page. One homepage for every comparison is how relevance dies. Our guides to ChatGPT ad landing pages and landing page message match cover the handoff. The message match checker is a quick pass before you launch.
Reusing comparison angles from Meta
Long-running Meta ads in your category often already use comparison framing: material explainers, "X instead of Y" hooks, side-by-side demos of your own products. The video will not fit a chat card. The decision criteria will.
- Find long-running ads in Meta's Ad Library, including your own.
- Write down the criteria the ad uses (cooling, width, size, ingredient, noise).
- Drop competitor names, rankings and anything you cannot prove.
- Rewrite the criteria as a title and a one-sentence body.
- Point the card at a page that shows the same criteria.
Our guide to reusing Meta winners in ChatGPT ads covers the rewrite in more detail. If you want a structured pass on a competitor ad's promise, proof and offer, the winning ad analyzer is built for that reading.
How to tell if comparison traffic is worth it
Comparison intent is not automatically profitable. A click from someone weighing three materials can still bounce if your price, shipping or product facts do not survive the comparison. Judge the traffic with your store numbers, not with a platform's reported return alone.
A practical read:
- Separate landing pages by question so you can see which decision actually converts.
- Tag links with UTMs so the source is visible in your store. Our UTM guide has a naming pattern.
- Watch new customers and margin, not only clicks. A page that wins on click-through and loses on returns or discounts can still be a poor use of spend. See new customer ROAS and contribution margin.
- Give a test enough time to include refunds if your category sees them. A same-week read can flatter a product people send back.
There is no universal benchmark for "comparison queries convert better." Run the test on questions you can answer with a real product, then keep the ones that produce customers you want after returns and margin.
A simple first test
- Pick three comparison questions you can answer with a specific product or variant.
- Write one context hint and one chat card per question. Run the copy through the policy checker.
- Build or tighten one landing page per question, with the criteria above the fold.
- Launch them as separate ads so you can read results by question. Our testing framework explains how to keep a test readable.
- After you have orders and a first look at returns, keep, rewrite or pause each question on its own.
Start narrow. Three honest answers beat a dozen cards that all say you are the better choice.
How SecondWin helps
SecondWin builds and runs ChatGPT ads for DTC brands. We read your site to see which products and variants you actually sell, then study the longest-running ads in your niche in Meta's Ad Library for the comparison criteria shoppers already respond to. We turn those into original chat cards, AI images and context hints aimed at "which should I buy" questions, and we preview the ad in a ChatGPT mockup before anything spends.
We do not write competitor attacks, invented rankings or statistics we cannot source. Campaigns run in your own OpenAI ad account. Get a free analysis of your store to see the comparison questions we would start with, or compare plans and pricing. SecondWin is an independent company and is not affiliated with OpenAI.
FAQ
Can I name a competitor in a ChatGPT ad?
It is a bad idea even when you are tempted by a "Brand A vs Brand B" question. Naming another company raises trademark and policy risk, and a "better than" line is a comparative claim you must be able to prove. Describe your product's facts and who it is for, and let the shopper compare.
Should my ad say it is the best option?
Only if you can substantiate that exact superlative. Most brands cannot. A specific fit ("percale for hot sleepers," "wide-fit daily trainer") is more useful to someone comparing options, and it is easier to support on the landing page.
What should the landing page include for comparison traffic?
Restate the decision in the first screen, show the criteria in plain language, compare your own variants fairly, and put shipping, returns and price where the shopper can see them. Send each comparison card to the page that matches that question.
How do context hints work for comparison questions?
Describe the decision the shopper is making, such as choosing between two materials or fitting a constraint like apartment size or wide feet. Avoid hints that name rivals or imply you know the private conversation.
Are comparison questions only for expensive products?
No. People compare low-priced products too, especially when the difference is material, ingredient, size or fit. The economics still have to work for your margin and your conversion rate. Use your own numbers, and see our ChatGPT ads cost breakdown before you scale a test.