Your strongest competitor went out of stock on Mercado Libre Colombia last night at 11pm. Your team finds out Monday at 10am. By then they've already restocked. How much did that 35-hour window cost you?

The four layers of retail media measurement — organic thermometer, organic×paid matrix, multivariate attribution, and the fourth one we're unpacking here — are built on the same foundation: consistent scraping of the digital shelf state. But the fourth layer has a feature the other three don't: it pays for itself every day, not in six months.

This layer is real-time operational scraping. Detecting competitor moves (stockouts, price changes, content changes, SOV ramps) and reacting with smart bidding, pricing adjustments, or campaign redirection within hours, not weeks. In this article we unpack the four most valuable types of operational signals and why combining them with the other three layers is what turns the measurement system into a complete retail media operating system.

The 4 operational signals that move ROAS

Four signals, in typical order of impact:

1. Competitor stockouts. When your strongest competitor goes out of stock, the searches they were going to capture redirect to the next product in the ranking. If you're well-positioned and raise the bid in that window, you capture traffic that wasn't yours under normal conditions. The window is usually short — 12 to 72 hours for high-velocity SKUs — but ROI during the window is disproportionate.

2. Competitor price changes. If a competitor raises price 8%, your product becomes relatively cheaper without you doing anything. That new relative competitiveness justifies raising the bid and pushing more traffic. If they lower price 8%, the opposite: lower the bid until you're not burning money on clicks with lower conversion probability against the new price.

3. Changes in competitor sponsored share-of-voice. If your competitor goes from 8% to 18% SOV in a week on an important keyword, you already know pressure is coming. The defensive action isn't waiting for your organic position to drop (that takes 4-8 weeks) — it's matching the SOV before and holding it. This signal connects directly to the organic×paid matrix from the previous article.

4. Content or product availability changes. New launches, hero image changes, listings appearing or disappearing from the retailer. These signals aren't as tactical as the first three, but they're the best way to detect new competitive threats before they show up in the thermometer.

The four together are the "combat radar" your retail media team needs to make well-directed daily decisions.

Why a competitor stockout is a window of gold (and how short it is)

It's worth diving into the first signal because it's the one brands most underestimate.

Imagine a competitive keyword — "ground coffee 500g." Five products rotate through the top 10 organic positions. One of them, the #2 competitor by share, runs out of stock on Mercado Libre Colombia on a Friday at 8pm. What happens?

In the first hours: the competitor's listing still appears, but with the "Out of Stock" button. Buyers who land on that listing (via organic or sponsored) don't buy there — they look for alternatives. Those redirected clicks go to the next best-matching product — usually the #1 or #3.

In the first 24 hours: the ranker detects the competitor's listing is out of stock and starts degrading its organic position for that keyword. That frees up even more visibility for the rest.

In 48-72 hours: if the competitor restocks earlier, much of that effect reverses. The listing returns, the ranker gradually restores it, the flow normalizes.

Result: during those 12-72 hours, the remaining products can capture 20-40% more traffic than normal on that keyword. The marginal ROI of the bid is enormously higher than under normal conditions — and almost nobody is taking advantage of it.

Why? Because the information arrives late. The brand finds out about the competitor stockout when someone notices manually, or when it shows up in Monday's report. By then the window has closed.

The solution is availability scraping every 1-4 hours on key competitive SKUs, plus automatic alerts when a competitor stockout is detected. The action is simple: raise bid 30-50% on the keywords where that competitor was prominent, hold until restock is confirmed. The marginal ROAS of that window is typically 2-3x the average ROAS.

When to raise the bid and when to lower it: the operational rules

The heuristic rules that work best for operational bidding:

Raise bid when:

Lower bid when:

These rules aren't academic — they're standard practice for mature retail media teams. The difference is that most teams apply them manually, weekly, to a handful of keywords. The team with operational scraping applies them automatically, in real time, to thousands of keywords.

Why the retailer's weekly dashboard arrives late

There's a reasonable question at this point: "but doesn't the Mercado Libre Ads, Amazon Ads, or Walmart Connect dashboard give me almost all of that?" The short answer: partially, and always late.

Why late? First, retailer dashboards report your performance, not the competitor's. You know your CPC went up 20%, but you don't know if that's because competitor X bid aggressively — the cause is hidden. Second, dashboard cadence is typically daily or weekly with a 24-72 hour lag. For operational signals like stockouts, that's eternity.

Why partial? The set of signals retailers report never includes complete competitive data. They'll never tell you "competitor X has 80% of stock critical on these SKUs." They'll never give you the competitor's SOV curve per keyword. That information is only available to the measurement system the retailer offers — and almost never at the granularity you need to operate.

That's why operational scraping is independent of the retailer: capturing the digital shelf state from the buyer's side, not the retailer's. It's what a customer sees when they enter to search. That view captures stockouts, prices, content, and paid SOV without waiting for the retailer to show you and without depending on their goodwill.

The connection to the other 4 articles in the series

The operational layer doesn't work in isolation. It's the foundation that feeds the other three:

Seen the other way: if you've invested in building operational scraping for daily bidding decisions, you already have the data foundation for building the other three layers. The biggest investment in the retail media measurement operating system is data capture. Once it's in place, the four layers are built on top with no comparable additional effort.

That's why this is the layer that pays for itself every day: each well-executed bid decision in real time (raise when a competitor went out of stock, lower when they cut price) generates incremental ROAS that more than covers the scraping cost. And as a bonus, you have the data to build the other three analytical layers on top.

What ePerfectStore does in real time

ePerfectStore runs two products directly on this layer that the brand team can use from day 1:

Stockout monitoring detects and alerts when your own SKUs or competitors' SKUs go out of stock on Mercado Libre, Amazon, Éxito, Jumbo, Olímpica, and other retailers. Configurable cadence up to hourly for critical SKUs. Alerts can be email, Slack, or webhook to your DSP or bidding tool.

Price monitoring detects listed and promo price changes from competitors in near-real time. Same alert infrastructure, same cadences. The data lives in history for elasticity analysis, category-relative comparisons, and as direct input to the observational attribution model.

Specifically for retail media, ePerfectStore captures additionally: paid SOV per keyword (which sponsored slots each brand occupies, week by week), content score per SKU (completeness, images, video), review velocity, and rating. Together with price and stockouts, that completes the inputs for the four analytical layers.

In summary

Operational signal Typical action Window
Competitor stockout Raise bid 30-50% on keywords where they were prominent 12-72 hours
Competitor raises price Raise bid on shared keywords Until next price change
Competitor lowers price Lower bid (or adjust your own promo) Until normalization
Competitor raises paid SOV Match SOV on key keywords for 6-8 weeks Quarter
Your own stockout Pause bid until restock Until restocked

The retail media measurement operating system isn't a methodology — it's an infrastructure. And the foundation of the infrastructure is consistent operational scraping. Without that foundation, the four analytical layers are academic exercises. With that foundation, they're tools that translate into daily bidding decisions, quarterly budget allocation, and honest CFO conversations at year-end.

If you've made it this far, you've read all five articles: the framework guide, the organic thermometer, the matrix, the detective, and the race. Those five pieces are the closest thing to a complete retail media measurement operating system that exists today in LATAM. The invitation is to start with the scraping and let the analytical layers be built on top — not the other way around.

Sources

Want real-time alerts on stockouts and competitor price changes across LATAM? ePerfectStore.com captures the digital shelf hourly — and feeds your retail media campaigns with that data.

Back to Blog