Cohort-Based AEO Reporting: Tracking Citation Lift by Article Vintage

Why Flat Citation Counts Lie

A monthly citation count looks simple. Total citations went from 200 in March to 240 in April to 290 in May. The chart goes up. The team feels good.

The problem is that this view hides the question that actually matters. Are the citations being driven by new articles published this month, or by articles published 6 months ago that are still accruing? If new articles are driving everything, the program is buying citations at a constant rate, not building an asset. If old articles are driving the lift, the program is compounding and each new month of work is worth more than the last.

The two scenarios look identical in a flat report. They are radically different in business reality. One is a content treadmill. The other is a flywheel. Cohort reporting is the structure that distinguishes them.

What a Cohort Report Looks Like

A cohort report groups articles by their publication month, then tracks citations from each cohort across subsequent months. The result is a table where rows are publication months and columns are calendar months, with each cell showing the citations driven by that cohort during that month.

The shape that matters is the row trajectory. The January cohort that produced 22 citations in its first month grows to 84 by month 12. That growth is compounding. If every cohort follows roughly the same shape, the program is healthy. If newer cohorts produce fewer month-1 citations than older cohorts did at the same age, something has changed (content quality, query competition, or model behavior).

What Cohorts Reveal That Aggregates Hide

Three insights are visible only in the cohort view.

The first is compounding rate. By comparing how cohorts grow from month 1 to month 6 to month 12, you can quantify the multiplier. A healthy AEO program produces a 3 to 5x multiplier from month 1 to month 12 on a per-cohort basis. Below 2.5x suggests either weak content, a category with low search persistence, or a model environment that does not reward old content.

The second is cohort quality drift. If May's cohort produces fewer month-1 citations than January's cohort produced, something dropped in editorial quality or query selection. The aggregate view does not show this because newer cohorts get masked by the accumulating contributions of older ones. The cohort view catches it the month it happens.

The third is decay onset by cohort age. By tracking cohorts past month 12, you can see when decay starts and which interventions slow it. A cohort that plateaus at month 9 and starts declining at month 14 reveals exactly when maintenance work needs to begin. Without cohorts, the decay shows up as "growth slowing", which is the wrong diagnostic.

How to Build a Cohort Report

The data inputs are simple. For each article, you need the publication date and a monthly citation count. For each calendar month, you need to attribute citations back to specific articles. Most measurement stacks already capture the second part. The first part requires a clean record of when each article went live.

The build sequence is straightforward. Tag every article with its publication month at the time of publish. Run the citation measurement monthly on the full content stack. Group the results by publication cohort. Plot the trajectory of each cohort across the calendar months that followed publication.

The trap is incomplete data. If you start measuring partway through a program, you have no baseline for the older cohorts, and their early-month performance is missing from the report. The fix is to set up cohort measurement at the start of an engagement, even if it takes 6 months to produce a meaningful first chart. Catching up retroactively is rarely accurate.

Reading the Cohort Shape

The shape of a healthy cohort follows a recognizable curve. Month 1 produces a baseline of citations from the initial retrieval propagation. Months 2 and 3 see strong growth as models build authority signals and the article gets internally linked. Months 4 through 8 are the steepest growth phase, often doubling the month-1 number. Months 9 through 12 plateau at the peak. From month 13 onward, maintained articles hold position while unmaintained articles begin a slow decline.

The slope of the early months tells you about retrieval. A cohort with weak month-1 to month-3 growth signals either weak content structure or low retrieval propagation, often because the articles were not internally linked or schema was missing. The slope from month 3 to month 9 tells you about authority accumulation. A cohort that plateaus early at low numbers signals authority gaps in the broader content stack, not in the individual articles.

Reading multiple cohorts together reveals program-level patterns. If every cohort plateaus at the same level, you have a topical authority ceiling that no amount of new publishing will break through without entity-level work. If newer cohorts plateau higher than older ones, your domain authority is growing and each cohort benefits from the prior ones. This is the signature of compounding.

The Per-Cohort Cost of Citations

Cohort reports also enable a real cost-per-citation calculation. The flat view computes cost-per-citation by dividing total monthly spend by total monthly citations, which is meaningless because the citations were earned by articles that have been compounding for varying lengths of time.

The cohort view divides the spend of a single cohort by the citations that cohort produces across its full life. A January cohort that cost 15,000 dollars to produce and drives 84 citations per month by month 12 has a marginal cost per monthly citation of 179 dollars at that age. Three months later, with the same one-time cost and a higher citation rate, that number drops further.

This is the calculation that wins renewal conversations. It shows that AEO spend behaves like an asset purchase rather than a media buy. A media buy stops working when you stop paying. An AEO cohort keeps paying out for years if maintained.

What OnlyAEO Reports Look Like

The default reporting view we ship to clients is a cohort matrix, updated monthly. Each month we show the cumulative citations earned by every cohort published since engagement start. The same view shows mention rate, cited article count, and competitive share.

Alongside the matrix we ship three derived views. Cohort velocity, showing how fast each cohort is climbing. Cohort decay watch, flagging any cohort that has lost more than 8 percent of its peak in the past 60 days. Cohort cost efficiency, dividing the cohort's production cost by its trailing-3-month citation count.

Clients use these views in renewal conversations, in board updates, and in internal stakeholder communication. The flat view satisfies no one because it cannot prove compounding. The cohort view does.