A claim ladder is defined as the ordered set of statements a brand makes about a product, from the declared composition at the bottom through the mechanism and the expected outcome to the promise at the top. To research competitors in skincare, read them as ladders rather than as pages is the fastest way to see where each one stops, because the point at which a brand stops climbing tells you what it is prepared to be held to. Around 70% of carts are abandoned across retail, and in considered categories the causes cluster in unanswered questions [Baymard Institute, 2025]. This guide sets out five research passes, what each is for, and how to turn the output into facts and tests.
- Run five separate passes rather than one sweep of competitor websites.
- Read every product page as a claim ladder and note where the brand stops.
- Score ingredient transparency first, because it is objective and predictive.
- Map the routine architecture: the sequence is the unit of sale in this category.
- Read the subscription terms in full and record where each one is disclosed.
- Convert every finding into a published fact or a test, or discard it.
Omniconvert has measured how commercial pages present evidence across the CROBenchmark dataset of 7,000+ websites in 15+ industries, against 248+ audit criteria, over 13 years in eCommerce [CROBenchmark Report 2026, Omniconvert]. Skincare stands out for the gap between how much a brand says and how much of it can be checked, which is exactly what makes structured competitor research productive here.
Why five passes beat one sweep when you research competitors in skincare
The familiar version of this work is a spreadsheet with a row per competitor and twenty columns of impressions. It takes a week, it is interesting to read once, and nothing changes because no column maps to a decision somebody was about to make.
Five narrow passes avoid that. Each one asks a single question across every competitor, which makes the comparison sharp and the output short. It also means you can stop after two if the first two answer what you needed.
The general framework behind this, across categories rather than inside one, is the DTC competitor research stack.
Pass one: map the claim ladder
Most skincare pages make four kinds of statement and most readers experience them as one blur of copy. Separated out they become comparable. One brand declares composition and stops. Another describes a mechanism but promises no outcome. A third promises an outcome and attaches a timeframe to it.
The top of the ladder is where regulatory exposure and commercial ambition meet, so a brand that climbs high is telling you something about its risk appetite and its evidence. A brand that stays low is often the one with the most careful customers.
Record it as four short cells per product rather than as prose. The pattern across a category usually becomes visible after about eight products.
Pass two: score ingredient transparency
| Pass | The question it answers | Effort |
|---|---|---|
| Claim ladder | What will they be held to | An hour per competitor |
| Ingredient transparency | Who expects to be checked | Minutes per competitor |
| Routine architecture | What is actually being sold | An hour per competitor |
| Subscription offer | Where the margin is defended | Minutes, if you read it all |
| Review recency sample | What is going wrong now | Half an hour per competitor |
Score three things: is the full list published as text rather than as an image of the carton, are active concentrations stated where the market permits, and are exclusions listed explicitly. Three yes or no answers per competitor, and the spread across a category is usually wider than anyone expects.
This pass is worth running first in practice even though it sits second in logic, because it takes minutes and immediately sorts the set into two groups that behave differently in every later pass.
The specific trap in skincare competitor research: a team collects screenshots of competitor product pages and compares them on design, because design is the thing a screenshot captures. The decisive differences in this category are an ingredient list published as text rather than as a photograph, and a subscription term disclosed before rather than after the commitment, and neither of those is visible in a screenshot of the hero section.
Pass three and four: routine architecture and the subscription offer
Start the third pass by trying to buy a complete routine from each competitor as a new customer. Note how many decisions you are asked to make, whether the order of application is stated anywhere, and whether the site assembles the sequence for you or leaves you to do it.
A brand that sells sequences captures a larger first basket and a much longer relationship. That shows up in site architecture rather than in copy, which is why it is missed by research built on page screenshots.
The fourth pass is quicker and more revealing. Record the subscription interval, the price after the first order, the cancellation route and, crucially, where each of those is disclosed. A 5% retention lift can raise profits by 25 to 95% [Harvard Business Review, 2014], so the subscription mechanics are the single most valuable thing on a competitor's site to understand properly.
AliveCor used Omniconvert to run a structured A/B testing programme and achieved +21% conversion rate, +5% revenue per visitor, and 94% statistical relevance across their experiments. [Omniconvert, AliveCor case study] That is a testing result rather than a research result, and it is the right end point for this work: competitor research that does not become a test is an opinion with footnotes.
Pass five: read the recent reviews, not the best ones
Sort by date rather than by rating, read the most recent thirty on each competitor's two biggest products, and tally the objections. The tally is short and stable, and it almost always names something you could publish or fix within a fortnight.
Pay particular attention to objections about things that are not the product: delivery, subscription changes, packaging, support response. Those are the gaps a competitor is least likely to notice, because their own reporting treats them as operations rather than as conversion.
Shopify's merchant guidance has pushed complete and consistent product information across channels for years, and a competitor's review corpus is an unusually honest audit of whether they have done it. The vertical sibling to this piece is how to research competitors in fashion, where the constraint is fit rather than suitability, and the independent research worth reading alongside all of it is at Baymard Institute.
What to do this week
- Score transparency across six competitors. Three yes or no answers each. Thirty minutes, and it sorts the field immediately.
- Buy a routine from the two best of them. As a new customer, noting every decision you are asked to make and every term disclosed late.
- Tally recent review objections. Thirty reviews per competitor, sorted by date, counted rather than read for sentiment. Then turn the top two findings into one published fact and one test.
Once you know where you stand, the free leaderboard entry at the competitor research stack is the quickest way to see your own six-dimension score against real competitors in your category and country. Downstream of that, Omniconvert Commander (formerly Nexus) is an AI for eCommerce growth engine that unifies commerce data, prioritises experiments by True Profit, and generates campaigns and creative you approve before they go live.
Frequently asked questions
How do you research competitors in skincare?
In five passes rather than one sweep: map the claim ladder, score ingredient transparency, reconstruct the routine architecture, read the subscription offer in full, and sample the review corpus by recency. Each pass answers a different commercial question, and doing them together produces a tour of competitor websites rather than a finding.
What is a claim ladder?
The ordered set of statements a brand makes about a product: the declared composition at the bottom, then the mechanism, then the expected outcome, then the promise at the top. Reading competitors this way shows you where each one stops, and the point at which a brand stops climbing is the most informative thing on its product page.
Why start with transparency rather than with positioning?
Because transparency is objective and positioning is not. Whether a full ingredient list is published as searchable text is a yes or a no, and it predicts how a brand performs with the most careful segment of this category. Positioning is worth studying second, once you can see who is willing to be checked.
How much does the routine matter in competitor research?
It is the unit of sale in this category. A competitor who sells a sequence captures a larger basket and a longer relationship than one who sells products, and the difference shows up in the architecture of their site rather than in their copy. Map the sequence before comparing any individual product.
Which reviews should we actually read?
The most recent ones, not the best or the worst. Recent reviews describe the product as it is now and the service as it currently runs, and the objections that recur across them name the gaps a competitor has not closed. Sorting by rating tells you about outliers.
How do we turn the research into something actionable?
Convert each finding into either a published fact you are missing or a test you could run. A finding that becomes neither is an observation, and observations accumulate without changing anything. Two facts and one test per research pass is a realistic and useful output.
Two facts and one test
Run the transparency pass across six competitors this afternoon, because it takes half an hour and it sorts the field into brands that expect to be checked and brands that do not. Everything you learn in the other four passes reads differently once you know which group each competitor is in.
Then hold yourself to a short output: two facts you should publish and one test you should run. Competitor research that produces a document produces nothing. Competitor research that produces three items on a backlog has paid for itself.