Reporting FAQs

Searches Report FAQs

Does this section display real-time data?

No; the query that powers this visualization does not take time into account. It only displays data from the specified days/dates.

Why does the search term not match the search summary table below?

1. This section displays the individual search terms while the table summary combines similar queries for those still using Beacon 1.0. Merchants using the new Beacon 2.0 search report will see search terms split by query; they are not combined.
2. The "most recent searches" are arranged in descending order of date, and the search summary table is arranged in descending order of count for combined queries.

For example, the search term "green" appears in the most recent searches but not in the summary table. This is likely because in the selected date range, a shopper must have searched for "green". But there are probably not enough counts of the search term "green" to make it appear in the immediate list within the summary section.

You can verify the results by using the filter function. In this example, the total count on the search term "green" is only 12, so it probably is sitting at the bottom of the table. This is expected because the search terms are arranged in descending order of the count.

Why do we have a new Searches 2.0 Report?

The older Beacon 1.0 search report lacked data points that allow you to see and interpret the search journey on your site. Engagement and conversion metrics were not available in the 1.0 version. The Beacon 2.0 initiative allows us to track and report on these data points, giving a better understanding of a shopper's journey.

The report is built so it can be used by both the buyer and ecommerce manager persona. It follows the traffic, engagement, conversion, and revenue flow in a shopper's journey.

Additionally, the new Beacon 2.0 API and tracking make it more intuitive for merchants to implement and for their teams to debug. We have streamlined the following as a result of Beacon 2.0:

• Adding specific endpoints to send data by the event
• Reducing data and requiring strict validation prevent bad data from being sent and enforces front-end issue resolution
• Providing elaborate error response messaging helps pinpoint the exact issue to resolve for both Athos and merchant teams to resolve.

How is this report different from the legacy search report?

Search queries are separated and reported on individually. Impression, click-through, conversion, and revenue metrics are now available. We report on all searches that hit the search API — which includes blank/empty searches.

Blank/empty searches happen when a shopper goes to the search bar and hits <Enter>. In this case, it loads the entire product catalogue (global campaign). We log this in our reports because merchants can see these interactions happen on their site, which gives them the opportunity to update their search strategy or, if they wish, to update the front end to filter these searches out. A high number of certain activities (e.g., empty searches) allows the customer to take action.

What powers the Searches 2.0 Report?

Shopper interactions on the front end are tracked and sent to the new Beacon 2.0 API.

What new data points are available in the Searches 2.0 Report?

• Impression rate
• Click-through rate (CTR)
• Average click position
• Add to cart
• Conversion rate
• Revenue

How is revenue attribution calculated?

If a shopper goes to the search bar, enters a search, and hits <Enter>, this counts as a search operation. If the shopper clicks on the search results and ends up buying a product within the 24-hour window, then that transaction is attributed to search. We use the last-touch attribution model.

What does session mean in our context?

Session ID (session_id) is an event that we receive from the merchant in their incoming data, so we do not generate it. This value can differ for each site based on their expiration logic and their setup for sessions.

Experiments Report FAQs

What happens when an experiment is set up? What happens on the Athos side?

When an experiment is run on a merchandising campaign and the variation is viewed, we look for userId; this is part of the tracking that should be configured. If this is not set up, it will by default return the first variation A. If you see only one variation getting views, it indicates that this userId is not being tracked correctly or is not being tracked at all.

The search API will take the userId and will randomly assign a variation of the campaign. This is then logged and recorded for the report. There is an A/B report updating job that runs to collect this A/B report data. We look for the variation ID of the experiment from the search beacon event. We count the number of times it has been viewed by userId (unique userId only). Next, we look at the transaction table for the same userId to check if they made a purchase.

Why focus on unique users?

Focusing on unique users is a guardrail against volume metrics. By showcasing unique users, you get a better sense of the distribution of engagement on the variation. It also more clearly defines the true size of the audience.

Since we only count unique users, this means that if the same user looks at the variation more than once, we do not log that, but other platforms may (such as GA4).

Which is better?

It depends on the specific goals and use case of the customer.

1. Overall engagement: Total number of views (volume) may be better to show how often a piece of content is consumed to understand popularity or stickiness
2. Audience Reach: Unique views are better, as they tell you how many different people your content is reaching

Note: In the legacy report, if the purchase is made within 24 hours, it is counted. If it falls outside the 24-hour window, it is not included.

New reports do not have a 24-hour criteria and are processed more frequently. What used to be excluded in the older system is now included, which means the numbers are slightly higher.

Is a view considered "unique" for the entirety of the test? If I, as a user, see a variation of the test on day 1, come back the next day and see it again on day 2, how is this tracked in the report?

Since we count the views per day, the above example should show a count of 2, if the user sees the tests on different dates.

Why do we see differences in numbers between GA and SS?

There will always be a difference when looking at numbers in Google Analytics and Athos. This is because of differences in tracking and the way data is reported.

How does currency selection and allocation work for the Experiments Report?

We updated this report on July 31st, 2025, so if no currency info is present, we update that value to the default currency (configured in website settings) This ensures that we do not lose those records. Thus, when the report loads, it will create that currency association in the background and load the report for the default currency selection. (Note that the currency selector at the top of the report always loads the default.)

Google Analytics Report FAQs

Do customers need to set anything up to get GA4 to work?

Yes, first and foremost, merchants need to set up Google Analytics 4 on their site. Our integration requires that setup. To collect the new GA4 data within Athos, merchants must follow a few steps to enable the right tracking and set up two audiences in their Google Analytics Console, which we will fetch using our API. These steps require no coding or technical knowledge to perform, and we have detailed instructions available.

Can a customer have a GA4 set up before Athos launches GA4 support?

Yes; in fact, Google presently advises customers to use Universal Analytics and GA4 simultaneously, which we are in line with. Google has expressed that these two tools do not interfere with each other and, within Athos, having both set up allows both Athos and our customers to sanity-check their GA4 setups against the UA data, to confirm consistency.

Will any data be lost in the GoogleAnalyticsConversionData document filter used to merchandise on things like ga_unique_purchases?

No, our testing has included the fields brought in by the GoogleAnalyticsConversionData document filter, which are all working in GA4.

I do not see purchase events show up when configuring an audience set up. What can I do?

We handle the purchase event based on the search data; the purchase data may not show up without these two settings:

1. Enable enhanced measurement and Site Search with corrected query parameter https://support.google.com/analytics/answer/9216061?hl=en&ref_topic=13367566&sjid=12138902835447054879-NA
2. Add e-commerce events to your websites https://support.google.com/analytics/answer/12200568?hl=en

Purchase event data is not showing up even though the audience is configured correctly

Check the query parameters in the Enhanced measurement section; Site search may have parameters with special characters. Also do a search and check what their query parameter is.

For example : https://wilton.com/search/?nav-user-quick-search=cake+decorator. Here, the query parameter has dashes (-), which are not allowed by  Google.

In this case, customers should reach out for support if this change needs to be updated with their Athos Integration/White Glove.

Refer to this Google doc for more info.

In the Query Parameter field, enter the word or words that designate internal query parameters, such as term,search,query,keywords. Sometimes query parameters are designated by just a letter, such as s or q. Enter up to five parameters, separated by commas. Do not enter any additional characters. For example, if the query parameter is designated by the letter q, enter only q (not q=) "



The customer has set up custom events to track search terms; how do we connect the audience?

We recommend using automatically collected Google events. However, there might be exceptions. If a customer has set up entirely custom events, it is likely that they may not be able to use the "view_search_results" in audience configuration. In this case, they can create the audience and select a custom event that replaces "view_search_results" . Additionally, an admin will need to manually enter the audience name in the Google Analytics settings page in the Athos console.

We cannot connect our Google account in the Athos console; are the Google analytics and audience configuration correct?

An error can occur if one of the audience filters contains different data types, e.g., count returning decimal instead of integer. This error occurs on the Google API, so there is nothing we can do to resolve this issue unless the client corrects their audience setup or we can manually add the audience in the Google Analytics settings page.

My Google Analytics Insights Report and search performance on the Athos console dashboard are not showing data despite GA4 configurations and SS audience created correctly. What can I do?

The Google Analytics Insights Report displays metrics comparing the "Athos" audience with the "All Users" audience, which is a default audience automatically generated by GA. Check to see if this default audience exists; if it does not, then that could be why you are not seeing the data.

The merchant will need to create this "All users" audience manually:

1. Go to the Audiences page > New Audience
2. Name the audience "All Users"
3. Under the "Include users when" section, choose the "session_start" event
4. Set the Membership Duration to "Maximum Limit"

Please note that creating this audience does not add data retroactively.

Category Report FAQs

Why did we create a new category report?

• In our legacy report, metrics for the category page such as page view, avg time on page, and bounce rate were sourced from GA4.
• The legacy category report lacked engagement and conversion data.

To address this and report more accurately, we added new endpoints that allow us to track these interactions from search and browse actions. This is important because merchandisers want to know how their PLP/categories are doing.

How is the new category report different from the previous one?

All data points are collected via front-end tracking and our Beacon 2.0 API.

How is revenue attribution calculated?

If a shopper goes to a category, clicks on the category results, and buys a product within the 24-hour window, that transaction is attributed to the category. We use the last-touch attribution model.

Autocomplete Report FAQs

How do we report on autocomplete?

We capture keystrokes, and the corresponding render is core to understanding user behavior. Whether the shopper completes the preliminary search or not, the interaction gives valuable insight into what is resonating or potentially failing in the autocomplete suggestions. This allows merchandisers the opportunity to fine-tune their autocomplete via autocomplete setup and merchandising campaigns to improve their site search experience.

What is included in the Autocomplete Report?

• Impression
• Number of results displayed
• Products clicked
• Click-through rate (CTR)
• Total filters

Autocomplete query columns, when sorted, can help the merchandising team see what the preemptive search action looks like.

General report FAQs

Which reports are updated/migrated?

• Zero Results Searches
• Searches
• Filters
• Spell Correction
• Experiments
• Product Recommendations

What has changed?

Search-related reports will be powered by ThoughtSpot.

Important: Merchants will be able to view historical data back 13 months for all reports.

What action is needed at launch to switch reports?

No action is needed; the reporting team made the switch, and the reports were updated automatically.

Why is the reporting view stacked, instead of a dashboard?

This is likely due to your screen size. Ideally, you should see the dashboard view. If you submit a support ticket for this issue, please include information about what browser you are using and your screen size.

What should I do if the Search, Filters, or Sorts Reports are not loading?

• Try refreshing; sometimes it could take a couple of seconds
• Submit a support ticket to request assistance; as a last resort, Athos can switch you to the old (legacy) report

Why is there a scroll bar within the table/chart?

This is inherited from the Thoughtspot user interface. If you are working on a bigger screen, these may be less apparent. In table summaries, this is unavoidable.

Will new reports get stuck like the legacy reports?

The new system should not have the same issues as we have seen in the legacy reports. However, we have a system in place to manually recover the data in a worst-case scenario.

Can I add an external email (that does not have ASD access) in the report schedule feature?

Yes, all users that have a login or account on the ASD will be able to schedule reports to an email they wish to add.

Can we export data from our reporting via API or web hook?

Download a PDF of the report by clicking the three-dot icon at the top right of the report and selecting "Download PDF". Tabular sections like Summary tables have an XLSX/CSV format available too.

Why does the popular query result not match results in the table in the Searches Report?

Popular queries give the popular search terms based on the date selected. This part of the report looks up individual queries, whereas the table at the bottom of the report combines similar queries. This is the reason the numbers or the order do not match up.

What value do new reports provide? / How do I use the "filter" function?

• New reports focus on shopper behavior tracking. To do this, we have removed the inclusion of background filters and only kept count of the ones shoppers have interacted with.
• New reports tie the search data throughout; drill down using the filter function to see how numbers change.
  • For example, on the Filters Report, right-click on the Search-Filter summary to use "Filter" functionality to get a drill-down of values selected in filters. Drill down on the search "lighting" shows two filters being used against it: merchantcategory and ss_price. Further down in the report, you can see the number of times a particular filter value has been used. In this case, "flush mounts" has the highest count, indicating that this type of product needs to be stocked more or can be optimized in a campaign.

Which reports use "sessionId"?

The reports using a sessionid include Searches, Filters, and Sorts.

Why do the numbers look higher in the new Experiments Report?

The older report had a 24-hour (one day) condition in place when displaying results, so it would take 24 hours to display any new information or update in numbers. The new report updates more frequently and is not bound by a 24-hour window.