Market Benchmarks

What are Market Benchmarks?

Market Benchmarks let you compare your own-store products against commercially similar competitor products. This is useful when exact SKU overlap with competitors is low, for example for private-label ranges, because it measures market position against comparable alternatives rather than identical items.

Benchmarks are deliberately separate from Exact Matches. An Exact Match means the same physical product found at another retailer. A Market Benchmark is a curated comparison choice: “would a shopper reasonably consider these two products as alternatives?”

Key concepts

Comparison Sets

Every benchmark is organized as a Comparison Set. A Comparison Set anchors on one of your own-store products and holds a collection of competitor products evaluated against it.

Comparable Candidates

A Comparable Candidate is a competitor product that has been evaluated or proposed as a benchmark for your anchor product. Each candidate carries a similarity score, shown as a percentage, and a status:

Status	Meaning	Counts as included
Suggested	Proposed by the generator; awaiting review	No
Auto-selected	Generator confirmed the candidate met the equivalence rule’s required attributes and confidence threshold; included immediately	Yes
Accepted	Approved by a reviewer, or selected as an Equivalent Match	Yes
Manual	Manually added by a user; protected from automation	Yes
Stale	A generated candidate superseded by a newer suggestion run; needs re-review	No
Rejected	Excluded; retained to suppress repeated suggestions	No

An included candidate counts toward the workspace summary cards immediately, but it only feeds price analytics for its anchor product when it is also that competitor’s Equivalent Match. See below.

Equivalent Match

An Equivalent Match is the closest commercial alternative to your anchor product at one specific competitor. Because each tracked competitor is evaluated separately, a Comparison Set can hold more than one Equivalent Match, up to one per competitor that has a benchmark candidate.

For each competitor, the Equivalent Match is whichever included candidate, whether accepted, auto-selected, or manually added, you’ve explicitly selected. If you haven’t made an explicit selection for that competitor, Aperture automatically falls back to its highest-ranked included candidate.

Market Comparables

Market Comparables are all other included candidates in a Comparison Set, once each competitor’s Equivalent Match has been set aside.

By default, price grids, price history, ranks, and competitor price position compare exact matches only. When switched on, the Exact + Equivalent match mode, described on Dashboard, folds in each competitor’s Equivalent Match wherever that competitor has no exact match for the anchor product; a product page shows an “Equivalent prices excluded” badge in Exact-only mode whenever the product has comparable candidates. Even with the mode on, only each competitor’s single Equivalent Match is used; other Market Comparables never feed price calculations. Market Comparables provide broader context on the review screens, not in price analytics.

Note

Equivalent Matches and Market Comparables are distinct from Common Products. Common Products are exact-match only by default. The Exact + Equivalent match mode reaches further than that one view. Switching it on also backfills a competitor with no exact match, with that competitor’s Equivalent Match, on Dashboard and its competitor statistics, and on a product’s price grid, price history, and ranks views on Product Detail; on price grids, Common Products, and the product’s Prices tab those substituted prices are marked approximate (≈). Benchmark decisions can therefore influence what appears across these views, but they never change the underlying exact-match data. See Matches & Match Review for detail on Exact Matches.

The Market Benchmarks page

The Market Benchmarks page provides a workspace-level view of all Comparison Sets.

Summary cards

Four cards appear at the top:

• Products with Benchmarks. Number of Comparison Sets that contain at least one Comparable Candidate. A Comparison Set whose candidates have all been deleted is not counted.
• Equivalent Matches. Total active Equivalent Matches across all Comparison Sets, counting each competitor’s Equivalent Match separately.
• Approved Market Comparables. Total included candidates, whether accepted, auto-selected, or manual, that are not an Equivalent Match.
• Needs Review. Total suggested or stale candidates waiting for a decision.

Comparison Set table

The table below the summary cards lists every Comparison Set. You can search by product name or brand using the search field above it. A line under the table shows how many Comparison Sets match the current search and filters out of the total.

The review badge on the Match rules page opens this page filtered to that category, with the needs-review filter on. Each active filter appears as a chip above the table, with an x button to clear it. A category filter covers the chosen category and everything beneath it. The needs-review filter keeps only Comparison Sets with suggested or stale candidates.

Column	What it shows
Anchor Product	Your own-store product the Comparison Set is built around
Product Area	The internal grouping area the Comparison Set belongs to. This is “default” for every Comparison Set created through the app, distinct from the anchor’s Category
Brand	Brand of the anchor product
Category	Category of the anchor product
Equivalent Matches	The anchor’s current Equivalent Match: the competitor product name when there is exactly one, “N retailers” when several competitors each have one, or ”—” when there is none
Included	Total included candidates: accepted, auto-selected, or manual
Needs Review	Badge count of suggested + stale candidates; click it to open the review panel
Last Updated	When the Comparison Set was last changed

Clicking a row always opens the product’s Matches tab on Product Detail. To open the inline review panel instead, click the Review button in the Needs Review column; it appears only when the Comparison Set has suggested or stale candidates.

Reviewing candidates inline

The review panel opens with an Equivalent Matches summary for the selected Comparison Set: the current Equivalent Match for each competitor that has one, or “Missing” if none exist yet. Below it, a status-filtered table lists the Comparison Set’s other candidates. The filter defaults to Suggested; switch it to also view Accepted, Manual, Stale, Rejected, or All. For each candidate:

• Suggested candidates show Accept and Reject buttons.
• Stale candidates show a Reaccept button.
• Any candidate can be removed with the delete icon. This permanently removes it from the Comparison Set.

Closing the panel after making changes automatically refreshes the overview table.

The Matches tab

The Matches tab, part of Product Detail, appears on a product’s detail page and groups comparable candidates by competitor, so you can review and select an Equivalent Match one competitor at a time.

• A row is shown for every tracked competitor. Competitors that already have an Exact Match for the anchor product carry no comparable candidates, because exact identity stays with the Matches & Match Review pipeline. Their row shows the “No product match under the configured match rules for this retailer” message.
• For every other tracked competitor, up to three candidates are listed. The current Equivalent Match and manually added candidates come first, and the rest are ordered by score. A competitor with no qualifying candidate also shows that “No product match” message.
• A header count shows how many of the listed competitors currently have an Equivalent Match selected, e.g. “4 / 9 retailers matched.”

Each candidate row shows its price at that competitor, flagged when out of stock, along with its score, up to three attributes it shares with the anchor product, and whether it’s the competitor’s current Equivalent Match. Available actions:

• Select as Equivalent Match. Makes the candidate the Equivalent Match for its competitor. Selecting a Suggested candidate also accepts it. If that competitor already has an Equivalent Match selected, you’re asked to confirm before replacing it.
• Remove from candidates. Deletes the candidate from the Comparison Set entirely.
• Report as incorrect. Flags the candidate for review without changing its status.

The Equivalents tab

The Equivalents tab summarizes every tracked competitor for the anchor product in a single table: whether the competitor has an Exact Match or an Equivalent Match, the matched product, its current price and availability, equivalent quality as a score, status, and when it was last updated. The product name and the equivalent-quality score appear only on Equivalent Match rows. Exact Match rows show the price, availability, and last-seen date only. Above the table, counts show the total number of tracked competitors alongside how many have an Exact Match, how many have an Equivalent Match, how many still have candidates awaiting review, and how many have neither.

• Manage opens the Matches tab so you can review or change that competitor’s Equivalent Match.
• Review appears for a competitor whose candidates are still awaiting review, alongside Add. It opens the Matches tab.
• Add appears for a competitor with neither an Exact Match nor an Equivalent Match yet. It opens the add-relationship wizard pre-filled for that competitor. Rows backed by an Exact Match show no action.

Adding a relationship manually

To add a competitor product as a benchmark, open the Equivalents tab and click Add next to the competitor you want to benchmark against. This opens the add-relationship wizard pre-filled for that competitor.

Choosing a relationship type

Select the relationship type before staging products:

• Market Comparable. The product is stored as a Manual candidate. Multiple products can be staged and added in one session. If the competitor has no explicitly selected Equivalent Match, a Manual candidate outranks accepted or auto-selected candidates and becomes that competitor’s Equivalent Match.
• Equivalent Match. The product is added with a perfect score and is included immediately. Only one product can be selected at a time; staging a second product replaces the first. It becomes the competitor’s Equivalent Match unless that competitor already has an explicitly selected Equivalent Match. In that case, use Select as Equivalent Match on the Matches tab to switch it over.

Manually added candidates count as included immediately. No review step is required. They enter price analytics only under the same rule as any other candidate: when the Exact + Equivalent match mode is switched on, and only if the candidate is that competitor’s Equivalent Match.

Searching and staging

Search for competitor products by name, brand, SKU, or EAN; the search term is not matched against the store name, since the wizard already scopes results to the competitor you chose. Click + on any result to stage it.

When you click +, a side-by-side confirmation dialog opens comparing the root product and the candidate. If EAN codes do not match, or one is missing, the dialog shows a warning. You can still proceed, but manual verification is advised.

Staged products appear in a highlighted list above the search results. You can remove any staged product before confirming.

Confirming the addition

Once you have staged your products, click Review (N) to proceed to the confirmation step and save the relationships.

Automatic suggestion generator

Note

The Aperture team runs the suggestion generator for your workspace. The Generate Suggestions button is not shown to customer users, including workspace admins. Everyone can still review and act on the candidates the generator creates.

The Aperture team starts the generator from the Generate Suggestions button on the Market Benchmarks page. The generator compares own-store products against competitor products in the same category, using deterministic scoring to identify commercially similar pairs.

The generator excludes pairs that share the same canonical product, GTIN, or MPN. Those belong to the Exact Match pipeline, not the Comparable Engine.

How suggestions are scored

The generator scores candidate pairs using these signals:

• Same category
• Shared product-type terms in normalized product names
• Normalized title token overlap
• Similar price band
• Compatible bundle quantity or net content where available

The generator then checks each scored pair against the equivalence rules configured for its category on the Match rules page, under Configuration → Match rules. It drops pairs that fail a required attribute entirely and counts them as skipped; they are never created as candidates. Of the rest, pairs that meet or exceed the rule’s confidence threshold, 70% by default and adjustable per category, become Auto-selected. The generator creates the remainder as Suggested for manual review.

The generator caps each generation run at 20 candidates per anchor product.

Running the generator

A dry run is required before you can write suggestions: Create N Suggestions stays disabled until a dry run returns at least one candidate, and it disables again after a run creates suggestions, until you dry run afresh.

1. Click Dry Run to preview candidates without writing anything.
2. Review the metrics and the preview list, up to 25 pairs, each showing a confidence percentage and reason.
3. If the results look suitable, click Create N Suggestions to write the candidates to the workspace.

Running the generator again later re-evaluates the whole workspace rather than piling up duplicates. It refreshes a previously generated candidate in place when the candidate still ranks among its anchor’s top matches, updating the score, status, and reason. It removes a candidate that no longer ranks, because a better match displaced it or the candidate no longer passes the rule. A suggestion-generation run never touches rejected or manually added candidates. It never re-scores or removes accepted candidates, but if the run’s rule version differs from the candidate’s, it marks the candidate Stale for re-review.

Note

After creating suggestions, the Market Benchmarks table can take a moment to catch up while the new candidates finish syncing into the workspace.

Generator metrics

After a dry run or a full run, the dialog shows:

Metric	What it means
Would create / Created	New candidates this run writes, counting both auto-selected and needs-review candidates, after the per-anchor cap is applied
Auto-selected	Candidates that passed all required attributes in their equivalence rule and the confidence threshold. Included immediately
Would update / Updated	Existing suggestions that still rank in their product’s top matches and were, or would be, refreshed with a current score, status, and reason
Would retire / Retired	Previously generated suggestions that no longer rank in their product’s top matches and were, or would be, removed; manually decided candidates are never removed
Skipped (required gate)	Pairs dropped because they failed a required attribute in the anchor’s equivalence rule
Total Candidates	All pairs that passed scoring and the per-anchor cap across the workspace
Anchors Evaluated	Own-store products with enough source data to be evaluated
Anchors with Candidates	Own-store products that received at least one generated candidate
Skipped Existing	Pairs skipped because that anchor-target combination already carries a decision the generator does not own: a manual decision, whether accepted, rejected, or manually added, or a candidate created by another source. Generated candidates are re-scored instead and appear under Updated
Skipped Exact	Pairs skipped because they appear to be Exact Matches: same canonical ID, GTIN, or MPN

A per-category breakdown table also shows how candidates are distributed, including how many anchors, auto-selected, and needs-review counts fall in each category.

After the generator runs

Generated candidates arrive with either Auto-selected or Suggested status. Auto-selected candidates count as included immediately. They appear in the Included count and summary cards without review. They enter price analytics under the same rule as any other included candidate: only when the candidate is its competitor’s Equivalent Match and the Exact + Equivalent match mode is switched on. Suggested candidates appear in the Needs Review count and must be accepted or rejected before they count as included.

The generator retains rejected candidates as suppression records so it does not repeatedly propose the same pair in future runs.

Candidate lifecycle reference

From	To	Trigger
(new)	Suggested	Suggestion generator proposes the pair below the confidence threshold
(new)	Auto-selected	Suggestion generator proposes a pair that already meets the equivalence rule’s required attributes and confidence threshold
(new)	Manual	User adds via the add-relationship wizard
Suggested	Accepted	Reviewer accepts, or selects the candidate as an Equivalent Match
Suggested	Rejected	Reviewer rejects
Stale	Accepted	Reviewer reaccepts, or selects the candidate as an Equivalent Match
Accepted / Auto-selected	Stale	A suggestion-generation run supersedes the candidate, because the run’s rule version differs from the candidate’s
Suggested	Stale	A suggestion-generation run supersedes the candidate, because the run’s rule version differs from the candidate’s
Suggested / Auto-selected / Stale	(removed)	A suggestion-generation run no longer ranks the candidate among its anchor’s top matches

The generator excludes manually added candidates from the Stale transition and from removal.

Automation cannot overwrite Manual or Rejected decisions. A manually added candidate remains included regardless of subsequent generator runs. A rejected candidate remains rejected unless a reviewer explicitly acts on it.

What Market Benchmarks do not cover

• Exact Matches and Common Products. The separate matching pipeline manages these; see Matches & Match Review.
• Equivalence rule authoring. The rules that drive scoring are configured on the Match rules page, under Configuration → Match rules.