A scaling signal is defined as an observable behaviour that only makes sense if a creative is working. To track a competitor's ad spend you cannot read the budget, because nobody publishes one and every tool offering a figure is modelling rather than measuring, so you log behaviour instead: live variants per concept, first-seen dates, survival, placement spread and message consolidation. The CROBenchmark dataset covers 7,000+ websites in 15+ industries, scored against 248+ audit criteria [CROBenchmark Report 2026, Omniconvert]. This guide shows how to log five signals weekly and read them together.
- Read committed behaviour instead of estimating a budget, because spend is never disclosed.
- Count variants per concept, the clearest public evidence of investment, since duplication is expensive.
- Check survival after a fortnight, because what is still live is what they believe in.
- Observe weekly on a fixed day, since everything here is a trend.
- Trust a reading only where advertising, hiring, landing pages and discounting point the same way.
Tracking competitor ad spend is a job that begins with accepting it cannot be done. No advertiser publishes its budget, no platform discloses another account's investment, and every tool that shows you a dollar figure for a private competitor is showing you the output of a model built on assumptions you cannot inspect. Last updated: October 2026.
Budget was never the real question. A team wants to know whether a rival is committing to something, and commitment leaves visible traces. Omniconvert has measured how stores read their competitive landscape across the CROBenchmark dataset of 7,000+ websites in 15+ industries, against 248+ audit criteria, and the teams who learn something record the same things on the same day every week.
This article covers what is observable, how to log it and how to read weak signals together. See the DTC competitor research stack for the wider survey, and how to Reverse-Engineer competitor ads with the Meta ad library for the creative itself.
Related reading: find which competitor ads are actually working.
How to track a competitor's ad spend: what you can and cannot learn
The platforms say as much themselves. Meta’s own Ad Library research tools documentation documents what the library withholds for commercial advertisers, and Google’s Ads Transparency Center publishes a comparable archive for Search and YouTube with the same absence of spend.
The five scaling signals that are actually observable
Variant count per concept. The strongest of the five. Producing eight versions of one idea costs studio time, editing and approvals, and no team does that for a concept still under evaluation. When variant count rises sharply on one idea, budget has been committed to it.
Creative survival. Note the start date of each creative and check a fortnight later which are still live. A creative running for months is a winner being exploited, and it tells you what their audience reliably responds to.
Placement spread. A concept that begins in one placement and appears across several has usually passed an internal threshold.
Format spread. The same argument appearing as static, as short video and as a longer edit means the concept has been judged worth adapting. Adaptation is expensive in a way that duplication within one format is not.
Message consolidation. A competitor whose creatives argue six different things is searching. One whose creatives all argue the same thing has decided, and what they decided is now readable in plain sight.
How to build the weekly observation log
- Fix the watchlist and the day. Five competitors is enough and ten is too many to sustain. Pick one weekday and keep it, because the value comes from even spacing rather than from coverage.
- Count live variants per concept. Group creatives by the argument they make rather than by their visuals, then count versions in each group.
- Record start dates and survival. New creatives this week, and which of last month's are still running. Survival is where testing separates from exploitation.
- Track placement and format spread. A short list per concept.
- Cross-check against non-advertising evidence. Job postings, landing-page changes, discount depth, delivery promises. A scaling read is only trustworthy when at least two independent sources agree with it.
Reading the signals together
The scaling ladder is defined as the ranking of public signals by how much of the competitor's own money each one commits.
| Signal | What it costs them | What it implies | How much weight to give it |
|---|---|---|---|
| A new creative appears | Very little | They are testing something | Almost none on its own |
| Creative still live after two weeks | Ongoing budget | It cleared their internal bar | Moderate |
| Variant count on one concept rises | Production and approvals | Committed budget behind that idea | High, the strongest single signal |
| Concept spreads across placements | Additional media investment | Passed an internal threshold | High |
| Concept adapted into new formats | New production, not just edits | Treated as a durable asset | High |
| All creatives converge on one message | Organisational agreement | Strategy settled, not exploratory | Very high, and easiest to miss |
The four mistakes that produce confident nonsense
Quoting an estimate as a fact. A modelled spend figure repeated twice inside a company becomes a number people plan against.
Reading heavy testing as heavy investment. A competitor running forty short-lived creatives may be spending less than one running four long-lived ones. Volume of creative and volume of money are different quantities and are frequently confused.
Sampling when you happen to remember. Irregular observation produces a comparison between a busy week and a quiet one, and there is no way afterwards to tell that from a real change.
Reading one platform as the whole picture. Social ad libraries are the most transparent source available and they are not the whole market. eMarketer's reporting on the growth of retail media shows how much advertising now runs in channels with no public library at all, so a competitor can scale substantially while your log stays flat.
What none of this tells you
It does not tell you whether the scaling concept is profitable, or what their economics are. Teams scale things for many reasons, including inventory pressure and simple momentum. You are observing a decision, not a result.
And it does not tell you where you stand. That requires measuring yourself on the same dimensions, which is what the six-dimension store score exists for: creative and ads, reviews and UGC, AI visibility, agentic commerce, competitor synthesis and CRO. A free leaderboard entry gives you the score, and the paid report gives you the working. Where the difficulty is that your own creative, order and margin data never sit together long enough to judge a concept, Omniconvert Commander (formerly Nexus) is an AI for eCommerce growth engine that unifies commerce data, ranks experiments by True Profit, and generates campaigns and creative you approve before they go live.
What to do this week
- Choose five competitors and write the names down. Direct rivals for the same purchase, not the biggest brands in the category.
- Take one baseline reading now. Variant counts per concept and start dates, today.
- Put twenty minutes in the calendar, weekly, on a fixed day.
- Add one non-advertising source. Their careers page is usually the cheapest and the most revealing, because hiring precedes spending and is published deliberately.
FAQ: tracking competitor ad spend and scaling signals
Can you actually see a competitor's ad spend?
No. No public source discloses what another advertiser pays, and any tool presenting a specific spend figure for a private company is presenting a model output rather than a measurement. What is genuinely observable is behaviour: which creatives are live, how many versions of each, how long they survive, and where they appear. Those are the inputs worth logging, and they answer the question you actually care about.
What is the strongest single signal that a competitor is scaling?
The number of live variants of one concept. Producing and running many versions of the same idea costs money and only makes sense once that idea is already working, so a concept that goes from two variants to a dozen is the clearest public evidence of committed budget. A single new creative appearing tells you almost nothing by comparison.
How often should I check a competitor advertising?
Weekly, on a fixed day, for a fixed list. The cadence matters more than the depth, because everything useful here is a trend rather than a snapshot, and a trend needs evenly spaced observations. Twenty minutes every Tuesday produces something worth reading in a quarter; two hours whenever somebody remembers produces a pile of impressions.
Do competitor ad spend estimates from tools have any value?
They are useful for direction and dangerous as figures. An estimate that says one competitor is investing far more than another is often directionally right, and the specific number attached to it carries error bars nobody shows you. Use them to rank, never to plan a budget against, and never quote the figure in a board pack without saying what it is.
What does it mean when a competitor kills a campaign quickly?
Usually that they are testing rather than that the concept failed, and the two are indistinguishable from outside. Short-lived creatives are the normal texture of an active advertising programme. The observation only becomes meaningful in aggregate: a competitor whose creatives all disappear within days is testing heavily, and one whose creatives run for months is exploiting a small number of proven ideas.
Should I copy a competitor concept once I can see it is scaling?
Copy the reasoning rather than the execution. A scaling concept tells you which objection or desire their audience responds to, which is genuinely transferable. The specific creative is tuned to their offer, their price point and their audience, so a close copy usually performs worse for you while also making your brand look derivative to anyone who sees both.
Most teams that try to track competitor ad spend end up quoting a tool estimate as a fact, and inside a company a modelled figure repeated twice becomes a planning number. A second failure is sampling when someone remembers, so a busy week is compared with a quiet one and no real change can be told from noise. Heavy testing also gets read as heavy investment, when forty short-lived creatives can cost less than four long-lived ones.
Before you follow a rival scaling concept, test it on your own traffic. 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]
Keep the boring log and check each reading against hiring
Give up on the number and you get something better than the number. A competitor's budget is private, unknowable and, once you have it, surprisingly hard to act on. Their behaviour is public, observable and directly actionable, because it tells you what they have decided to believe about their own customers. So keep a boring log. Five names, one weekday, five fields, twenty minutes. Count how many versions of each idea are running, note which creatives survive the fortnight, watch for a concept spreading into new placements and formats, and pay particular attention to the moment their messages stop diverging and start agreeing. Then check the reading against something outside advertising before you act on it.