A thermometer doesn't tell you which virus you have. It tells you whether you're getting better or worse. Your organic position in ecommerce is that thermometer for your retail media — and almost nobody is reading it well.

In the previous article — the honest guide to measuring retail media — we argued that the right question isn't "how much did each peso of my spend return?" (that question can't be answered well) but "are we winning the digital shelf?". That second question can be answered, and the main tool is weighted first-page organic position.

Sounds complicated. It isn't. But it's poorly understood in most marketing teams we talk to. Let's break it down step by step.

What "weighted first-page organic position" means in human language

Let's go piece by piece.

"Organic position" means where your product appears when someone searches, without anyone having paid for it to appear there. If I search "Greek yogurt" on Mercado Libre Colombia and your product comes up in fourth place — not in the sponsored ads at the top, but in the regular results — your organic position for that keyword is 4.

"First page" means you only count positions a customer actually sees. Studies consistently show that more than 80% of clicks in ecommerce land on the first 10-20 positions. If you're at position 247, your position exists technically but commercially you're invisible. That's why the thermometer is built on the first page, not on the universal ranking.

"Weighted" means not all keywords carry the same value. The keyword "milk" gets hundreds of thousands of monthly searches on Mercado Libre Colombia. The keyword "premium lactose-free milk 1L" gets a few hundred. If your brand moved up 5 positions on the first and dropped 3 on the second, that's an excellent week — but a simple average would treat it as a tie. Weighting fixes this: each keyword counts proportional to its search volume.

Put it all together: if you have a curated universe of, say, 500 keywords relevant to your category, you measure your organic position on each one week to week, and you average weighted by volume — you have a single number that summarizes "how well are you being found by your customers this week." That number, tracked over time, is your thermometer.

Why 6–12 months, not 6 days: the ranker has memory

Here's something important that almost nobody explains well. The thermometer doesn't work for short measurements.

Why? Because the ranking algorithms of Amazon, Mercado Libre, VTEX (which powers Éxito and Jumbo in Colombia), and practically every retailer have memory. They don't react to what happened yesterday. They react to accumulated trends over weeks or months.

Specifically, the variables that move the ranker most are:

All of these variables are slow. Sales velocity needs a few days to average. Conversion rate needs a minimum impression volume to be statistically significant. Reviews accumulate at a few per month for most products. Rankers compute these averages with rolling windows, then use the results to reorder search results.

Result: if you increase your retail media spend today, your organic position isn't going to move this week. Probably not next week either. But if the spend is well-directed, in 4-12 weeks you'll start to see the effect. And if you keep it consistent, you'll see a durable improvement that persists even if you lower the bid.

This means two things:

1. The thermometer is monthly, not daily.
2. The decisions you make based on it are quarterly, not weekly.

If you stare at the thermometer day to day looking for small movements, you'll get frustrated. The signal is in the multi-month curve, not in the daily noise.

The virtuous circle (and why it isn't magic)

There's a circle worth making explicit because many teams don't understand why organic position ends up being a good reflection of paid spend effectiveness:

Sponsored ad appears in a prominent position → some customers click (the ad is where the eye goes) → some of those clicks buy (because the product is relevant) → those clicks and conversions get registered by the ranker → the ranker concludes the product is popular for that keyword → the product climbs in organic → now the product also receives organic clicks and sales → those reinforce the ranker → and the position holds.

This circle is academically documented. A paper published by Amazon's ranking team in 2016 — Sorokina & Cantú-Paz, "Amazon Search: The Joy of Ranking Products" — describes it explicitly: the ranking algorithm heavily weights engagement signals from recent impressions, and sponsored placement pushes those signals. A recent academic audit (Information Economics and Policy, 2024) confirms with at-scale data that sponsored products receive orders of magnitude more clicks than non-sponsored ones at parity of relevance, and that those clicks are measurably reflected in subsequent organic ranking.

That's why the organic thermometer is a reasonable proxy for retail media effectiveness: if your spend is well-directed, the circle closes and organic position reflects the success of the investment. If your spend is poorly directed (bid too low, irrelevant keywords, weak product offer), the circle doesn't close and position doesn't move. The thermometer detects both situations — and, most importantly, the difference between them.

Reviews: the most underestimated ranker input

Of the six variables that move organic position, reviews are the one brands underestimate most — and the one that can move the thermometer fastest when worked with discipline. A data point cited by Tom Cohley, CEO of Cohley, on the Digital Shelf Institute (September 2024): per BazaarVoice data, products with fewer than 50 reviews suffer a minimum 30% drop in conversion rate vs. peers with sufficient volume. The Nature's Way case documented in the same article shows a 20% conversion lift on Amazon after implementing an authentic UGC strategy.

Three nuances that matter when reading the thermometer:

When you see a thermometer drop and the organic×paid matrix doesn't explain it, check competitor review velocity. It's the second place to look — after competitor paid SOV.

What this looks like at each LATAM platform

It's worth zooming in on how this materializes at the retailers that matter in the region:

Mercado Libre (Colombia, Mexico, Argentina, Chile, Brazil). Here's an important specifics: the seller's "reputation thermometer" is a first-order input to visibility. Complaints, cancellations, and shipping delays during the last 3 months (or the full history if you have fewer than 40 orders) determine whether your product shows up high or gets penalized. Premium listings pay more commission and get a boost. Mercado Envíos Full (the FBA equivalent) is a clear advantage. The organic thermometer on ML has to consider all of this, not just raw position.

Amazon (LATAM and global). The algorithm (originally A9, now closer to COSMO) prioritizes keyword relevance, sales velocity, conversion rate, reviews, availability, FBA/Prime, and increasingly external traffic. The thermometer works well here because the algorithm is relatively meritocratic — position reflects product performance in the category.

Éxito, Jumbo, Olímpica (VTEX). These retailers run on VTEX, an ecommerce platform with a configurable search engine. The important difference: there's a layer of human curation on top of the algorithm. Category managers can pin products in specific positions, create curated "showcases," and run commercial campaigns that override the pure organic order. The thermometer works, but you have to read the composition — sometimes a product is at position 1 because it genuinely performs, and sometimes because the category manager pinned it. Distinguishing between the two is part of the art of reading the thermometer.

Walmart (Mexico). Algorithm closer to Amazon's, with special emphasis on "listing quality score" (attribute completeness, image quality), shipping speed, and seller scorecard metrics.

The thermometer is built, ideally, with separate coverage per retailer. A single number that averages all of them hides important differences — you might be winning on Mercado Libre and losing on Éxito at the same time, and the average tells you "all good" when half the strategy is broken.

The 3 thermometer traps

The thermometer isn't perfect. It has three known traps, and it's worth mentioning them so they don't catch you by surprise.

Trap 1: Keywords where you're already #1. If you've been the organic #1 for "anti-dandruff shampoo" for three quarters and nothing's going to move, the thermometer doesn't flag that as a problem — but it's a keyword where your sponsored spend is probably cannibalizing your own organic clicks. This is exactly the problem the next article in the series tackles, on the organic × paid matrix.

Trap 2: Positions that improve without increased sales. In highly fragmented categories with hundreds of similar SKUs and little differentiation, it's possible to climb 2-3 average positions without sales moving visibly — because the customer sees twenty similar products and the difference between position 7 and position 5 is marginal. The thermometer rises; the business doesn't. That's why the thermometer is always cross-checked with sales data to validate.

Trap 3: The "first-page cap." The thermometer measures up to position 20 or 30 (depending on the retailer). If you're already inside that range for all your important keywords, you don't have much room to climb. The signal flattens. In that case, the battle moves to the organic × paid matrix and operations (holding stock, defending price, winning content) more than to "moving up positions."

Knowing these traps doesn't eliminate them — it just lets you read the thermometer with humility. It is what it is: a vital sign, not a diagnosis. It tells you whether the fever rose or fell. To know what caused it, you need the other layers.

How the thermometer is built in practice

Operationally, building a useful thermometer has three requirements:

ePerfectStore.com automates these three steps across the main LATAM retailers (Mercado Libre, Amazon, Éxito, Jumbo, Olímpica, and others). The brand team sees the weekly weighted-position curve, broken down by the dimensions they need to make decisions — without having to build homemade scrapers, maintain integrations with each retailer, or settle for the limited dashboard each platform offers.

In summary

Aspect What you need to know
What it measures Average organic position, weighted by search volume, across first-page keywords.
Cadence Weekly capture, monthly read, quarterly decisions.
Why it works The ranker has memory; it reflects sales velocity, conversion, reviews, and price.
What it does NOT tell you Causal channel attribution; cannibalization on keywords where you were already winning.
Traps Saturated #1 keywords, improvements without sales, first-page cap.

If your retail media works, your weighted organic position will improve over 6-12 months. If it doesn't work, it won't move. It's a slow, honest thermometer that's reasonably hard to manipulate. It doesn't replace a geo experiment when the moment of truth comes, but it does answer — week to week, month to month — the question your CFO is asking: are we winning or losing on the shelf where our customers buy?

In the next article in the series, we'll explain why this thermometer alone can lead you to bad decisions — and how to cross it with your sponsored share-of-voice so the picture becomes much more useful. Read: Looking at your paid and organic position together: the trick that changes everything.

Sources

Want to see your organic thermometer this week? ePerfectStore.com captures weighted organic position on Mercado Libre, Amazon, Éxito, Jumbo, and more — week by week.

Back to Blog