A quarterly competitor audit is defined as the same set of scoring axes applied to the same set of competitors, four times a year, so that the change between quarters is readable. The cadence is the point. A one-off audit tells you where everybody stands, which is mildly interesting; four audits on fixed axes tell you who moved, which is actionable. Pick three competitors for different reasons and keep them for a year: the closest rival, the category leader, and one store doing something structurally different. Score coarsely, ahead, comparable or behind, with one line of evidence beside each. Then end with one sentence naming the single gap you will close, and discard everything else until next quarter.
- Fix the axes before the first quarter and never edit them. An edited axis destroys the comparison.
- Three competitors, chosen for different reasons, kept for a year. Six is twice the work and half the signal.
- Score ahead, comparable or behind. A ten-point scale implies precision you do not have from outside.
- Read the change against last quarter, not the level. The level is what a one-off audit already told you.
- End with one sentence naming what you will do. An audit with eleven findings and no commitment changes nothing.
Most competitor audits are run once, filed, and repeated eighteen months later against different axes by a different person. Last updated: October 2026.
Omniconvert has scored stores against their category rivals across the CROBenchmark dataset of 7,000+ websites in 15+ industries, against 248+ audit criteria, over 13 years in eCommerce, and the audits that change behaviour are not the thorough ones. They are the repeatable ones. Nexus by Omniconvert is the AI for eCommerce growth engine that unifies commerce data and ranks the next action by True Profit, and the discipline behind it is the one this article is about: the same measurement, applied the same way, often enough to show a direction. The dimensions themselves are in DTC competitor research stack.
Why the cadence matters more than the depth
It is worth being clear about what a quarterly rhythm actually adds, because thoroughness is the instinct and it is the wrong one.
A one-off audit tells you that a competitor has better product photography and a slower delivery promise. You suspected both. The document is circulated, agreed with, and filed.
Four audits tell you that the same competitor's delivery promise tightened by two days in the second quarter and their ad concepts stopped rotating in the third. Neither fact is visible in a single snapshot, and both are the kind of thing that precedes a strategic move.
Which means the axes are sacred. The moment somebody improves the scoring sheet between quarters, the comparison breaks and you are back to a sequence of unrelated snapshots. Keep a separate list of axis ideas and introduce them at a stated annual reset, with the old axes kept running alongside for a year.
Picking three competitors, and why not six
The selection is a one-hour decision that governs a year of the exercise.
The closest rival is the store your customers also considered. It is the most directly useful comparison and the least surprising, because you are probably already watching it informally.
The category leader may be far larger than you and that is fine. You are not comparing resources, you are reading the expectations they set, because a shopper who has been on their site judges yours against it whether or not that is reasonable.
The structurally different store is the one most teams omit and the one that produces the findings. A store selling the same thing through a wholly different model, subscription instead of one-off, marketplace instead of own site, has solved problems you are not solving at all.
Keep all three for a year. Swapping a competitor mid-year has the same effect as editing an axis, and the temptation arrives in the quarter where one of them has done nothing interesting.
Scoring coarsely, with the evidence beside it
The scoring discipline is where most audits quietly become unusable.
One axis has a genuinely good public source and is worth doing properly: creative. Meta publishes its Ad Library precisely so that anybody can see which ads a page is running and for how long, which makes the creative axis the one place where your outside view is close to complete. Record concepts live and days running rather than your opinion of the work, because longevity is evidence and taste is not.
Resist the ten-point scale. It feels more rigorous and it is less reliable, because the difference between a six and a seven on somebody else's creative strategy is not a thing you observed. Three values are enough to rank gaps and robust enough to compare across quarters.
Write the evidence next to the score, in one line, always. "Behind: their delivery promise is three days, ours is five" is a line somebody can check. "Behind on delivery" is an opinion, and in a meeting it will be met with another opinion.
Leave a cell blank rather than estimating. Where an axis needs internal data you do not have, the blank is the honest entry, and it also records that this axis is unobservable, which is itself worth knowing when you next review the sheet.
The table sets out which axes are observable from outside, what evidence each one needs, and the error each invites.
| Axis | Evidence available from outside | The usual error |
|---|---|---|
| Creative and ads | Concepts live, days running, format mix | Judging taste instead of longevity |
| Reviews and UGC | Volume, recency, whether replies exist | Comparing averages, ignoring recency |
| AI visibility | Whether assistants name them, and accurately | Asking once instead of on a fixed set |
| Agentic readiness | Machine-readable price, stock and returns | Checking the rendered page only |
| Delivery and returns terms | What the product page states, per market | Reading the home market and stopping |
| Conversion rate | Nothing. Leave it blank | Estimating it, then quoting the estimate |
The last row is the one to enforce. An estimated competitor conversion rate becomes a fact within about six weeks of being written down, and decisions get made on it.
Reading the diff, not the level
This is the step that justifies the previous three, and it takes twenty minutes.
Put the two sheets side by side and mark only the cells that changed. In a typical quarter there will be three or four, which is a short enough list to actually discuss.
Then ask what each change implies. A delivery promise that tightened means they changed carrier or warehouse, which costs money and signals priority. Ad concepts that stopped rotating can mean they found something that works or that the budget moved elsewhere, and which of those is true usually becomes clear next quarter.
Watch for a change on your own row too. The sheet should include you, scored by the same people on the same axes, and a quarter in which you moved backwards on an axis without deciding to is the most useful finding the exercise can produce.
On the AI visibility and agentic readiness axes, be careful to measure the same way each time. Asking an assistant once produces a different answer each session, so use a fixed prompt set, and check the machine-readable facts with scripting disabled rather than by reading the rendered page. Stripe's developer documentation on agentic payment flows is a reasonable external reference for what a readiness axis should look for, and Baymard Institute's checkout research is the standing external source for the delivery and returns axis.
What a growth team should do this week
The first audit takes most of a day. Every one after it takes two hours.
- Write the axes and freeze them. Six is plenty. Note the date you froze them and the date they may next be revised.
- Pick three competitors for the three stated reasons, and write down why each one was chosen so next quarter's reader knows.
- Run the first sheet with three values and an evidence line per cell, including a row for your own store, and blanks where nothing is observable.
- Write the one sentence. The single widest gap you can close this quarter, as a commitment with a name attached.
- Diarise the next one for the same week next quarter, and store the sheet where somebody who inherits the job will find it.
Where competitor observations, your own analytics and your product data have to sit together before a gap can be sized, that unification is what Nexus is built for, with every proposed change staying something a person approves before it goes live. If you want the comparison scored against real stores in your category and country rather than against three you chose, Ecommerce Benchmark will place you on the free leaderboard across six dimensions, including the AI Visibility and Agentic Commerce axes that most audits have no way to assess.
FAQ: running a quarterly competitor audit
How often should you run a competitor audit?
Quarterly is the right interval for most categories, because it is long enough that competitors have actually changed something and short enough that you can still act within the same planning cycle. Monthly produces noise and audit fatigue. Annually produces a document nobody compares with anything, since by the time the second one exists the axes have usually been rewritten.
Which competitors should a quarterly audit cover?
Three, chosen for different reasons and kept for a year. The closest rival, because it shares your customers. The category leader, because it sets expectations you are judged against. And one store doing something structurally different, which is where the surprises come from. Auditing six is a common error: it doubles the work and dilutes the comparison without adding a decision.
Why should the scoring scale be so coarse?
Because you are observing a competitor from outside with no access to their numbers, and a ten-point scale implies a precision you do not have. Three values, ahead, comparable and behind, are honest about the evidence and still enough to prioritise with. A finer scale makes the exercise feel rigorous while making the quarter-on-quarter comparison less reliable.
What should a competitor audit not try to measure?
Anything that requires their internal data: conversion rate, margin, customer acquisition cost, retention. Those get estimated, the estimate gets quoted in a meeting six weeks later as a fact, and a decision is made on a number somebody invented. Score what is observable from outside and leave the rest blank rather than modelled.
What is the output of a good competitor audit?
One sentence naming what you will do this quarter, plus a dated scoring sheet filed for comparison. The sheet is working notes and the sentence is the product. An audit that produces eleven observations and no commitment has cost a day and changed nothing, and it will be repeated next quarter with the same result unless somebody insists on the sentence.
The bottom line
Freeze six axes, pick three competitors for three different reasons, and score them coarsely with one line of evidence in every cell. Include your own store in the sheet, leave blanks where nothing is observable from outside, and refuse to estimate anything that needs their internal numbers. Then do it again in exactly thirteen weeks and read only the cells that changed, because the diff is the entire return on running this quarterly rather than once. Finish with one sentence naming the gap you will close and whose job it is. The sheet is working notes; the sentence is the product, and an audit without one has cost you a day.