🩺 Frequency Optimisation

Dashboard link: https://app.emax-digital.com/frequency-optimisation-from-amc

Overview

The Frequency Optimisation dashboard uses Amazon Marketing Cloud (AMC) data to answer one of the most practical questions in digital advertising: how many times should you show an ad to a shopper before performance starts to drop off? The dashboard groups users into "frequency buckets" and then measures how each group performs on key metrics like Purchase Rate, Add to Cart Rate, Detail Page View Rate, and Click Rate. This reveals the sweet spot where additional impressions still improve outcomes, and the point of diminishing returns (or even negative returns) beyond which budget is being wasted on over-served audiences.

This dashboard requires an active Amazon Marketing Cloud (AMC) integration. AMC is Amazon's privacy-safe data clean room that lets advertisers analyse cross-channel audience and conversion data at the user level — without exposing personally identifiable information.

How to read the rates and the buckets. Each chart plots one metric per frequency bucket. A frequency bucket groups users by how many times they were exposed to an impression — 1×, 2×, 3× … up to 25× (a user in the "3×" bucket was exposed exactly 3 times). Each rate is defined as that metric ÷ impressions within the bucket — e.g. Purchase Rate = purchases ÷ impressions, Detail Page View Rate = detail page views ÷ impressions. Buckets are AMC-defined and cannot be customised.

What You Can See on This Dashboard

Section: Purchase Rate by Frequency

At how many ad impressions per user does the Purchase Rate drop, and how many users are in each frequency bucket?

Key metrics:

• Purchase Rate — purchases ÷ impressions, calculated within each frequency bucket.

• Average Users in Bucket — how many unique users received that many impressions, on average, across the selected campaigns and period.

Visualizations: A bar chart showing Average Users in Bucket and Average Purchase Rate side-by-side for each frequency bucket. The bucket with the highest Purchase Rate marks your optimal frequency point; the drop-off beyond it indicates where additional impressions stop converting.

Section: Add to Cart Rate by Frequency

At how many ad impressions per user does the Add to Cart Rate drop?

Note: Add to Cart Rate is only tracked for DSP (Demand-Side Platform) campaigns. This section will show no data for Sponsored Ads-only campaigns, as Amazon does not record cart actions for Sponsored Ads clicks in AMC.

Key metrics: Add to Cart Rate (add-to-carts ÷ impressions) and Users in Bucket per frequency bucket.

Visualizations: Bar chart of Users in Bucket vs. Add to Cart Rate by frequency bucket.

Section: Detail Page View Rate by Frequency

The earliest meaningful engagement signal in the funnel — often peaks at a lower frequency than purchase, since detail page views require less commitment than a sale.

Key metrics: Detail Page View Rate (detail page views ÷ impressions) and Users in Bucket per frequency bucket.

Visualizations: Bar chart of Users in Bucket vs. Detail Page View Rate by frequency bucket.

Section: Click Rate by Frequency

Useful for diagnosing creative fatigue — if Click Rate falls sharply after a small number of exposures, your creative may be wearing out faster than the rest of the funnel.

Key metrics: Click Rate (clicks ÷ impressions) and Users in Bucket per frequency bucket.

Visualizations: Bar chart of Users in Bucket vs. Click Rate by frequency bucket.

Section: Aggregated Trends by Ad Product Type

A cross-campaign view, showing how frequency-driven performance compares across different ad format types (DSP, Sponsored Products, Sponsored Brands, Sponsored Display).

Key metrics:

• Average Unique Users per Frequency — average number of unique users reached at each frequency level, broken down by ad product type.

• Average Purchase Rates by Ad Product Type — purchase rate by frequency bucket, aggregated across campaigns and broken down by ad format.

Visualizations:

• Average Unique User per Frequency by Ad Product Type — line chart showing how many users sit in each frequency bucket per ad format.

• Average Purchase Rates by Ad Product Type — line chart showing how purchase conversion behaves across frequency buckets for each ad format, side by side.

Available Filters

Filter interaction notes

• Two independent filter sets. The four campaign-level frequency sections (Purchase, Add to Cart, Detail Page View, Click) use their own Marketplace and Report Date filters; the bottom Ad Product Type trend section has its own pair. Changing the "top" filters will not affect the "bottom" section, and vice versa.

• The Campaign filter scopes only the four campaign-level frequency sections. The aggregated trend section at the bottom is driven by Ad Product Type and is not campaign-specific.

• Reporting Range and Report Date work together. Switching from Weekly to Monthly reloads the available Report Date values.

• Campaign Name search is a free-text filter that complements the Campaign dropdown — useful when you maintain naming conventions (e.g. BRAND_DE_RETARGETING_*) and want to grab a whole campaign family at once.

Use Cases

1. Find your purchase-rate sweet spot — Select a key DSP retargeting campaign, set Reporting Range to Weekly, and read the Purchase Rate chart. The frequency bucket where Purchase Rate peaks is your optimal frequency cap to set in the DSP campaign manager.

2. Detect over-serving and reclaim budget — If Purchase Rate drops sharply after, say, 4 impressions but you still have large user counts at higher frequencies, your campaign is over-serving a saturated audience. Lower the frequency cap and reallocate the freed budget to reach more first-time viewers.

3. Compare DSP vs. Sponsored Ads efficiency — Use the Ad Product Type filter in the aggregated trends section to compare whether DSP or Sponsored Display hits its purchase-rate peak at a lower frequency. A lower-frequency peak means more efficient conversion per impression.

4. Diagnose creative fatigue — Compare the Click Rate chart against the Purchase Rate chart for the same campaign. If clicks fall off well before purchases peak, your creative is wearing out — refresh or rotate assets. If clicks stay strong but purchases drop, the issue is on the detail page or pricing, not the ad.

5. Optimise Add to Cart pacing for DSP retargeting — Use the Add to Cart Rate section to understand at what frequency shoppers are most likely to add to cart — often earlier than the purchase-conversion peak. Set a tighter frequency cap on upper-funnel placements and a higher one on retargeting.

6. Evaluate seasonal changes in frequency sensitivity — Compare the same campaign's frequency curve across different reporting periods (e.g. peak season vs. off-season). Shoppers may tolerate higher frequency during high-intent periods like Q4 or Prime Day.

Limitations & Notes

• 14-day data lag. Data is always fetched 14 days after each reporting period's end, so Amazon's standard 14-day attribution window can complete. Results for a given week or month are available on the 15th day after that period closes.

• Attribution model. Conversion metrics use AMC's standard 14-day click + 14-day view attribution window, which may differ from the attribution settings in your campaign console.

• Add to Cart Rate is DSP only. The Add to Cart section returns no data for Sponsored Ads campaigns — Amazon does not track cart actions for Sponsored Ads clicks within AMC.

• AMC privacy thresholds. AMC enforces aggregation minimums to prevent re-identification of individual shoppers. Frequency buckets with very small audience sizes may be suppressed (shown as zero or empty) — this is an AMC requirement, not a data error.

• Two separate filter scopes. The campaign-level sections and the Ad Product Type trends section have independent Marketplace and Report Date filters. Always confirm both are set to your intended period before drawing conclusions.

• Per-bucket rates, not cumulative. Each bucket represents users who were exposed exactly that many times — not "at least" that many. A drop at the 5× bucket means users exposed exactly 5 times converted worse than users exposed 3 or 4 times, in isolation.

• AMC integration required. This dashboard will be empty without an active AMC connection. Contact your emax digital account manager if data is missing.

Data Refresh

Data updates daily, with a 14-day lag after each reporting period ends. For weekly data, complete results for a given week are available on day 15 after that week closes.