Skip to main content
Use Insights to get a daily, plain-language read on what is moving in your storefront. Each insight describes a change Layers detected in recent search and purchase activity, why it matters, and what you can do about it. Insights are useful when you want a single place to scan for movement without building reports.

Prerequisites

  • Your store must be connected to Layers.
  • Your catalog should be synced.
  • A few days of shopper search and purchase activity must exist so Layers has a baseline to compare against.

Open Insights

  1. In the Layers dashboard, open Lab.
  2. Select Insights.
The page opens with a table that lists every insight. Before your first insights are ready, the page explains that they arrive after the next nightly run. Layers needs about 7 days of search and order data to compare against. Select Choose insight types to pick which categories you want while you wait. See Choose which insights Layers looks for.

Find insights in the table

Each row shows:
  • Verdict: How to treat the finding. See Verdicts.
  • Insight: The headline describing what changed.
  • Subject: The query, collection, product, or other item the insight applies to. When the subject has a Performance panel, select it to open the panel over the period the insight covers.
  • Status: Active or Dismissed.
  • Type: The kind of finding, such as Trends or Anomalies. See Insight types.
  • Period and Detected: The time window the insight covers and when Layers detected it.
Rows are grouped by verdict, with Needs attention first, followed by Opportunities, Watch, Working well, and For your info. To narrow the table:
  • Filter by Verdict to show only Problem, Opportunity, Watch, Win, or FYI insights.
  • Filter by Status to see active or dismissed insights, or by Type.
  • Select Search to search by headline or subject.
The filters stay next to Search above the table. Your search and filters are saved in the page URL, so you can share a filtered view with your team. Open the actions menu on a row to view the details, copy a link to the insight, run one of its recommended actions, or resolve it. Select Copy link to share a direct link that opens that insight’s details.

Verdicts

Every insight has a verdict that tells you what to do with it:

Insights while you work

You do not need to leave your current task to check relevant findings. Insights also appear on the Analytics page, one finding per section or category, with controls to browse them and compare current values with the baseline. Select See why to open the finding’s details, or View all to open the full Lab feed. Cards show comparison bars or revenue charts when the finding includes that data. Each insight uses its own displayed timeframe; changing a dashboard’s reporting date range does not recalculate the insight. Dismiss hides a card on that page for you in the current browser; Undo restores the last dismissed card. If you dismiss every card, select Show dismissed insights to bring them back. Dismissing a card does not resolve the finding for your team or stop automatic recovery. Use Resolve or Not relevant in the details panel for those actions. Recommended actions follow the same review flow as the Lab feed. In particular, a sort-order action opens a staged edit; it does not publish a change automatically. See Apply an insight to a sort order. If no section appears, there may be no relevant active findings or you may not have permission to view them. Ask your administrator to check access to relevancy insights if needed. Sort-order actions also require permission to edit sort orders.

How the feed is organized

Each insight has a type, shown in the Type column and available as a table filter:
  • Trends: Search demand moving up or down. Use these to spot terms gaining or losing interest before they show up in headline metrics.
  • Attribution: Revenue influenced by search and discovery. Use these to see which queries, collections, or products are driving the biggest share of search-led revenue. Attribution cards highlight wins, so a window where attributed revenue dropped is suppressed in favor of the up window for the same subject. Genuine revenue declines surface as anomalies instead.
  • Anomalies: Breaks from your recent baseline in store-wide search click-through or engagement. Layers runs two checks: a sharp break, where the latest complete day falls far outside your recent daily baseline, and a slower trend, where the last 7 days average well below (or, for click-through, above) the 28 days before them. Trend cards catch gradual declines a single-day check would miss, and they’re only raised when the change is both statistically clear and at least 10%. Use these to investigate issues before they affect more shoppers. This category also flags storefront surfaces with no requests for 14 days.
  • Opportunities: Demand you are not fully capturing yet, such as high-traffic queries with low engagement or searches shoppers ignore. This category also holds sort health insights that flag mis-windowed sort orders, collection grid insights that flag top sellers, pins, and products held back or misplaced in your busiest collections, and setup cleanup insights that flag unused audiences and stale merchandising rules, and new and returning visitor gaps that flag collections where first-time shoppers engage much less than returning ones.
  • Sequence suggestions: Signals about your product sequences based on how often their members are actually purchased together. Layers surfaces two kinds of sequence cards:
    • Products bought as alternatives: Members of an existing sequence rarely appear in the same order and one product clearly dominates the others in sales. Shoppers are treating them as alternatives, not a set. Open Product Sequences to consider splitting the sequence or moving the leader first. Layers only raises this card when the dominance gap is large and the trailing product has enough orders to be a reliable sample. A normal 2:1 popularity split between styled-together items will not trigger it.
    • Product pairing: Two products are frequently purchased together but are not in any sequence yet. Consider creating one in Product Sequences.
    Non-merchandise items (for example, tip, gift-wrap, and shipping-protection line items) are excluded so pair suggestions stay focused on real products. Pair titles are sorted alphabetically, so “A & B” and “B & A” collapse to a single card.
  • What Layers did: A collapsed section that lists AI curation decisions applied to broad queries. When AI reranking reorders a high-volume query’s front page, a card lists how many products were promoted or demoted, the rationale, and the specific products that were demoted. Expand the section to review the demoted products and retag any that were misclassified. Cards only appear here when a demotion occurred or Layers recorded a rationale for the decision.
If the same subject (for example a query term like “anklet”) would otherwise produce several near-duplicate cards, Layers collapses them to the single highest-signal card so the feed stays scannable. Conversion-rate “converts at” cards are also capped per page so a single class of insight cannot crowd out the others.

Choose which insights Layers looks for

You can turn off kinds of insights your team doesn’t use. You need the Configure store settings permission.
  1. Go to Settings → Data & Sync → Insights.
  2. Review each category. The table shows a description, an example insight, and how many insights in that category are open now.
  3. Turn a category’s switch off to stop Layers from looking for that kind of insight, or on to start again.
The categories are Search demand, Search performance, Revenue from search, Alerts, Collection grids, Sort orders, Product sequences, Setup housekeeping, and AI curation log. Install health is always on and can’t be turned off, because a broken storefront integration can cost sales before anyone notices. Turning a category off hides its open insights and stops new ones. Changes apply from the next nightly run. Insights you already resolved stay in your history.

What each card shows

Every insight card includes:
  • A badge with a plain-language label, such as Sales spike, Search shoppers ignore, or Rule ready for review. Its color tells you how to treat the finding, independently of the category it appears in.
  • A percent change versus the baseline when a comparison is available. On sharp-change cards, the change indicator follows the finding’s badge color rather than always showing an amber warning.
  • A timeframe row with the time window the insight covers (for example, 7d), the period label, and the date Layers detected the change.
  • A headline describing what changed.
  • The subject the insight applies to, such as a query cluster, collection, or product group.
  • The current value for the underlying metric.
  • Comparison bars showing the current value against the baseline, or a daily revenue chart on attribution cards with a daily breakdown.
  • A short explanation of what changed and why it matters.
  • A recommended action when Layers has one, such as reviewing a query, adjusting a sort order, or adding a merchandising rule.
For a search with no results or only a few matching products, start with Test Text Search and check product availability in Catalog. Review Search Instructions if the AI needs clearer merchandising guidance. Layers does not have a manual synonym editor or a screen called Search Tuning. Use the badge colors to decide what needs your attention:
  • Red: An active problem affecting your store, such as a storefront block receiving no requests or a sharp drop in search engagement.
  • Amber: A fix to review, such as limited product coverage, low engagement, a search shoppers ignore, or a sort that uses a poorly fitting sales window.
  • Green: An opportunity, such as a sales spike, growing demand, or a strong seller that shoppers find too far down a collection.
  • Gray: A setup check, such as an unused audience or a rule ready for review, or an automatic curation update. These are not urgent problems.
A card in Anomalies is not necessarily a problem: Sales spike has a green badge because it can be an opportunity. Likewise, setup checks in Opportunities use gray badges. Read the badge and recommended action together before changing your merchandising. Your weekly report email uses the same badge colors. It groups findings into Opportunities, Needs attention, Setup notes (at most two), and What Layers did for you. Each emailed finding includes a specific next step, Open in Layers, and Learn more. For a finding whose only action is checking an installation, Open in Layers opens the relevant installation guide.

Open a finding and take action

Select a row in the table to open the insight’s details. The details view leads with the evidence behind the finding, a one-sentence summary, and the recommended actions. When the subject has a Performance panel, select Open performance panel to see its full performance over the insight’s period. Below that, expand the analysis to read what changed, why it matters, the recommended action, and what Layers has already done. When an automatic fix exists, Fix status shows its progress and any available evidence, such as the affected search, the search it borrowed results from, or why a fix was rolled back. Available actions depend on the finding:
  • Test in search opens the search testing page with the affected query already filled in.
  • Review semantic redirects opens Semantic Redirects for supported search coverage findings. Use a redirect when a search should lead to a specific page, not to fill a gap in product results. Review query expansions opens Query Expansions when you have access; this is an admin-only settings screen, not a synonym editor. Neither action creates anything automatically.
  • View product opens the affected product in your catalog.
  • Sort health actions open a staged edit for you to review, as described in Apply an insight to a sort order. These require permission to edit sort orders. Actions that add a metric-based boost or switch metrics also require Edit analytics, because Layers adds the metric to your store if it doesn’t exist yet.

Resolve a finding

If you have permission to manage insights, the choices depend on the verdict:
  • Problems and Opportunities: Choose Resolve when you have addressed the finding, or Not relevant when it does not need action. In the row actions menu, the second option is called Mark not relevant.
  • Wins and FYI: Choose Got it to acknowledge the finding.
  • Watch: These findings can’t be resolved. They clear on their own once the signal returns to baseline.
To bring back a dismissed insight, filter the table by Status → Dismissed, open the row’s actions menu, and select Restore. Resolving or marking a finding as not relevant hides the grouped finding from the feed and keeps it hidden through routine refreshes. It can return if the underlying finding changes materially.
Resolving or marking a finding as not relevant also stops automatic recovery for that grouped finding. It does not apply the recommended change or undo a fix already applied. Review Fix status before dismissing it.

Read the revenue chart

Attribution cards with a daily breakdown distinguish loaded revenue from days that are still loading. A dashed line may continue the recent trend for up to three loading days. The dashed line is a visual projection, not final revenue or a revenue forecast. Hover over a projected point to compare Projected trend (not final) with Recorded so far. Totals exclude loading days. When there is not enough recent data to extend the trend, the chart says so rather than projecting the remaining days.

Searches shoppers ignore

A Search shoppers ignore card flags a search that gets real traffic but whose results shoppers rarely click. The headline shows how many searches it got in the last 7 days, how often shoppers clicked a result, and your store’s typical click rate for comparison. Layers only raises the card when the search’s click rate is very low (5% or less), is no more than half of your store’s typical search, and the gap is large enough that it isn’t just chance. Searches that do get clicks set your store’s typical rate, so many ignored searches can’t hide each other. The amber badge marks a fix to review. Run the search in Test Text Search and review the first page of products. When a closely related search performs well, Layers can also try an automatic fix and report its progress in Fix status. These cards replace the earlier zero-result cards, because AI search rarely returns no results at all.

Anomaly review

When Layers detects a store-wide drop in click-through or engagement, it automatically re-checks the search results for the query patterns most likely to be dragging the metric down. Each anomaly card then reports one of two states:
  • Reviewed healthy: Layers re-ran the affected searches and confirmed the results look right. The drop is probably driven by external factors (seasonality, traffic mix, a specific campaign) rather than a search-quality problem.
  • Needs merchant action: Layers found the results for one or more affected searches look off. Run them in Test Text Search, then review the affected products or their rules in Merchandising Rules.
The anomaly review runs in the background once per day and never changes live ranking on its own. It only flags where to look.

Storefront surfaces with no requests

Layers watches whether each storefront surface it powers is still receiving shopper requests. When a surface that should be active receives no requests for 14 days, an Anomalies card such as Blocks has received no requests for 14 days. Check install. appears. This usually means the theme no longer calls Layers for that surface, for example after a theme change or a removed app block. Layers checks three surfaces:
  • Blocks: Checked when your store has at least one active block that has existed for 14 days or more. A new installation of Layers restarts this 14-day wait.
  • Search: Checked once your storefront has sent text or image searches to Layers since the latest installation.
  • Browse: Checked once your storefront has sent collection page requests to Layers since the latest installation.
Select Check install to open the setup guide for that surface: Rendering blocks in Liquid for blocks, Implementing search in Liquid for search, or the installation guide for browse. After the surface starts receiving requests again, the card clears on a later daily refresh. While Layers is still processing recent storefront activity, the card stays in place rather than clearing early. Layers only checks stores that have finished onboarding.

Sort health insights

Layers checks your published, metric-first sort orders every day and raises cards when the ranking is likely to be missing recent movement. These cards appear in Opportunities and Anomalies and always target a specific sort order or product, so the recommended action is a concrete change instead of a general note. Sort health insights include two types:
  • Sales spike: One card for your whole store, covering the last 7 days, that lists the products selling well above their usual daily pace. Layers looks at each product’s strongest day in the last 7 complete days and compares it with that product’s daily sales over the 28 days before. The card lists up to 10 products, strongest first, and the details show the units each product sold on its spike day. The card refreshes in place as the week moves on instead of adding a new card each day. You see a green opportunity badge, not a warning. Consider moving those products higher in the collections they appear in while demand is up.
  • Ranking window mismatch: The sales window configured on a sort order’s primary metric trails the window Layers found actually predicts next week’s sales best for your store. For example, a sort order using a 30-day window when a 7-day or trend-weighted window wins the backtest by a clear margin. The amber badge asks you to review the suggested change.
Layers runs a per-store sort window backtest once a week that compares trailing 7/14/30/60-day windows against 7-day and 14-day half-life decayed windows. When a short-horizon window wins, the Trending sort recipe and the Trending measures in the metric picker become available, and the window-mismatch insight can recommend switching to them. Stores without enough recent purchase history for the backtest to be reliable don’t get these insights.

Apply an insight to a sort order

Sort health cards, and some collection grid cards, carry one or two action buttons under the recommended action. Selecting an action opens the target sort order in the editor with the change already staged. Boost actions add the soft boost directly above the sort order’s main ranking metric. Metric switches replace that main ranking row, keeping any conditions it already had, instead of whatever happens to sit first. Nothing on the storefront changes until you publish. Available actions include:
  • Boost proven sellers: Adds a soft boost to the sort order that lifts products in the top 10% by 30-day revenue. Appears on Buried seller cards when the sort order is the cause.
  • Boost products that convert: Adds a soft boost to the sort order that lifts products in the top 10% by 30-day conversion rate. Appears on Clicks without orders and Deep clicks cards.
  • Switch metric: Swaps the sort order’s primary metric to a Layers recipe with a better-fitting window (for example, the Trending recipe or a shorter trailing-units window). The Switch to trending variant only appears when the sort window backtest supports a short-horizon window for your store.
  • Allow proven sellers past the family cap: Opens the sort order with Let top sellers past the cap enabled on the relevant diversity expression, using the default exception settings. Review and publish the change yourself; selecting the action doesn’t change your live sort order.
When you open the editor from an insight action, a Prefilled from an insight banner appears above the sort form. The banner reminds you the change is staged and not live yet. Review the staged expressions, tweak anything you want to adjust, then publish through the normal review flow. You can change a staged boost’s metric, its in top % threshold, or any other setting before you publish; see Boost by a metric. Dismiss the banner to hide it for the current session; the staged changes stay in the form either way.

Collection grid insights

Layers reads the product grid shoppers actually see on your highest-traffic collections, using each collection’s default sort and the merchandising rule that applies when no shopper-specific conditions match. It compares each position with the last 30 days of clicks and orders, then raises cards in Opportunities when pins, diversity caps, merchandising rules, or the sort order itself hold back products that sell. A collection qualifies when it gets a meaningful share of your store’s collection clicks over the last 30 days. The bar scales with your own traffic, so mid-size stores get these insights on their main collections, not just very large stores.
  • Buried seller: One of your store’s top sellers by 30-day revenue gets plenty of clicks on a collection but sits well below the first page, or outside the products Layers loaded for the collection. The card names the most likely cause:
    • The collection’s diversity cap pushed the product out of the top window.
    • A demote expression in the collection’s merchandising rule pushed it down. The card names the rule when it can.
    • Most of the products ranked above it are pinned.
    • Nothing else is holding it back, so the sort order’s metrics place it low.
    When a diversity cap is the cause, the card shows how many positions the seller lost when that information is available, along with the family members holding the in-window slots. If the seller falls outside the loaded grid, you may not see a positions-lost count. Use Allow proven sellers past the family cap, when offered, to review a bounded exception in the sort editor. When the sort order is the cause, the card offers Boost proven sellers and, when your store supports it, Switch to trending units (7-day half-life). See Apply an insight to a sort order. For pins or demotions, review the collection’s merchandising rule.
  • Dead pin: A pinned product draws a lot of clicks on a collection but converts far worse than the collection’s typical product. The card shows the pin next to the best-converting unpinned product near the top of the collection as a suggested replacement. Pins always win, so update the pin in the merchandising rule if you agree. Layers raises at most one card per rule, led by the most-clicked dead pin.
  • Clicks without orders: Several unpinned products near the top of the collection get many clicks but convert far worse than the collection as a whole. The card lists better converters further down the collection and offers Boost products that convert so products that convert can hold the first page.
  • Deep clicks: Shoppers click deep into a collection instead of the first 12 products. A card such as Only 18% of clicks in Summer Dresses land in the first 12 slots appears when less than 30% of the collection’s clicks land in the first 12 slots, or when its share is below three quarters of your store’s median across collections. The headline shows the store median when Layers has one. Layers only checks collections with at least 10,000 clicks spread across at least 12 positions in the last 30 days. The card lists up to five of the most-clicked products sitting past slot 12, or below the products Layers loaded for the collection. Pin the products shoppers look for, or use Boost products that convert to review a soft boost that lifts high-converting products in the collection’s sort order. Edit merchandising rule opens the collection’s merchandising rule when it has one; otherwise, Open merchandising rules opens your list of rules.
These cards describe the default sort with no shopper context. Collections that change for specific audiences, or that shoppers usually view with a different sort, can look different on your storefront.

New and returning visitor gaps

Layers compares how first-time and returning shoppers respond to the product grid on your busiest collections. When new visitors engage much less than returning visitors on the same collection, an Opportunities card such as New visitors have much lower CTR than returning visitors on Summer Dresses appears with a 30-day timeframe. Layers checks two rates for each collection:
  • CTR (click-through rate): How often a collection page view leads to a product click.
  • CVR (conversion rate after a click): How often a product click leads to an order.
A card appears when either rate for new visitors is below 60% of the same rate for returning visitors. Layers only checks collections with at least 20,000 collection request impressions in the last 30 days, so small samples don’t produce cards. When both rates qualify, you see one card for the collection, led by the larger gap. The card shows both rates side by side for New visitors and Returning visitors, along with the collection’s request impressions for the same period. A gap like this often means the collection’s default order relies on signals that matter more to returning shoppers. Consider giving first-time visitors their own sort order:
  • Create new-visitor sort: Opens a new sort order for the collection. The editor reminds you to target each ranking expression to the New visitors audience, and sets Preview as to that audience when your store’s New visitors audience still uses its default conditions. Preview the collection as a new visitor before saving. Creating the sort doesn’t change your storefront. To use it, assign it in a merchandising rule that targets new visitors.
  • View New visitors audience: Opens the audience so you can review who it includes. The button uses your audience’s current name, and only appears when your store still has its New visitors audience.
Creating a sort order requires permission to view and edit sort orders. Viewing the audience requires permission to configure the store and view merchandising rules.

Setup cleanup insights

Layers also checks your setup for leftovers that no longer do anything useful. You see these cards in Opportunities with gray setup badges and a 7-day timeframe. In your weekly email, they appear in Setup notes, with at most two notes and no setup note chosen as the priority highlight.
  • Unused audience: A saved audience isn’t used by any sort order, merchandising rule, or block. Attach it where you want to tailor results for those shoppers, or delete it if you no longer need it.
  • Rule ready for review: An enabled merchandising rule still pins several products but hasn’t been edited in months. Products pinned that long ago may no longer be the ones to feature. Rules on your busiest collections are listed first. Recent edits from the dashboard or an AI agent count as activity, so a rule you just reviewed isn’t flagged. Open the rule and confirm it still matches your current merchandising.
  • Rule off the default sort: A live merchandising rule uses a different sort from its collection’s default, so it only applies after shoppers choose that sort. Each card groups the affected rules for one collection, including rules with contextual conditions. Your busiest collections rank first. Preview rules and the All collection are excluded.
A rule off the default sort may be intentional during an A/B test. Layers skips affected rules when it can link them to a running Layers experiment. If you have a running ABConvert or Intelligems test, the note stays visible, names the test, and explains that the setup may be intentional. It has lower priority than other setup checks. Layers cannot confirm which rules those external tests use, so check your test setup before changing a rule or the collection’s default sort. For a rule not on the default sort, the card shows the collection’s resolved default and each affected rule’s sort. With merchandising-rule edit access for the collection, you can make a rule’s sort the collection default or move the rule to the current default. Both actions ask for confirmation. Changing the default affects everyone browsing the collection without choosing a sort, including which other rules apply. A move keeps the saved pins, expressions, and targeting. Layers blocks conflicting moves and gives you a link to the existing rule. Moves from an insight are also blocked while a shared editor or review is active; open the rule to make the change there instead. If the rule or default has changed since the card was generated, refresh before trying again. The card disappears after the next insight refresh once all affected rules are fixed. See Fix a rule that isn’t on the default sort.

Troubleshooting

The feed is empty

Layers needs about 7 days of search and order data before insights start appearing, and it looks for new insights each night. If your store is new or recently re-launched, check back after more traffic has accumulated. If you expected a specific kind of insight, check that its category is turned on in insight settings.

A card I expected is missing

Insights are filtered to keep the feed high-signal. Movements below the configured thresholds (for example, very low search volume or a small percent change) are intentionally hidden. Cards also drop off the feed once the underlying signal returns to baseline. Layers also collapses near-duplicates to the strongest card per subject, so you see one card per query term.

A surface shows no requests, but it works on my storefront

Open your storefront in a private browser window and use the surface, for example by running a search or opening a collection page. If results load from Layers, the card clears after requests reach Layers and the next daily refresh runs. If the card returns, use Check install to compare your theme with the setup guide, or contact Customer Care.

An anomaly stays in Needs merchant action

The state updates after the next daily review. If the affected query patterns still look weak the next day, the card stays flagged until either the results improve or the anomaly itself clears.

A resolved finding returns

Routine refreshes keep a resolved finding hidden, but a material change can bring it back. Open the details to review the new evidence before resolving it again.

I cannot resolve a finding or open a sort edit

Resolving findings and opening sort edits depend on your permissions. Ask your store administrator to review your access. If an action fails, refresh and try again; contact Customer Care if it keeps happening.

Next steps