How to Measure Search-Attributed Revenue

"Search drove 12% of revenue last month" sounds like a clean number. It isn't, until you know what it's measuring. Search-attributed revenue can mean the last click before checkout came from a search result, or that a session touched search at any point, or that a specific recovered query led to an add-to-cart within a time window — and each rule produces a different figure from the same underlying data. Before a number like this goes into a dashboard or a board deck, the methodology behind it matters more than the percentage itself.

What Is Search-Attributed Revenue?

Search-attributed revenue is the share of purchases, cart activity, or overall revenue credited to a shopper's interaction with search, as opposed to browsing, recommendations, or another discovery path. It sits under the broader concept of search attribution — the methodology a store uses to decide how much credit a purchase gets for having involved a search interaction in a session that likely touched more than one path.

There's no single industry-standard method. That's the part teams skip past, and it's the part that makes two stores' "search drove X% of revenue" claims impossible to compare unless both state their rule.

Why the Attribution Rule Matters More Than the Number

The same session data can produce very different revenue-share figures depending on a handful of decisions:

Last-touch vs. any-touch. Does a purchase count as search-attributed only if search was the last interaction before checkout, or any time search was used in that session? Any-touch rules produce higher, softer numbers.
Session window. How long after a search does a resulting add-to-cart or purchase still count as attributed — the same page view, the same session, or a multi-day window? Shorter windows are more defensible but attribute less.
Query match strictness. Does the add-to-cart need to tie back to the exact query text, or just to any product viewed after a search in that session?
What counts as "search." Autocomplete selections, zero-result recovery, and typed queries that return results are three different interactions. Lumping them into one "search" bucket hides which one actually drives the number.

None of these choices is wrong on its own. The problem is reporting the output number without the rule that produced it — a 12% figure and a 26% figure might reflect the same underlying shopper behavior measured two different ways.

A Practical Framework for Measuring It

1. **Define the attribution event before looking at data.** Pick one rule — for example, "an add-to-cart or purchase in the same session, within a fixed time window, where the product was reached via a search result or a search-recovered result." Write it down before running the numbers, not after seeing which rule produces the better headline. 2. **Separate full search from recovery.** A shopper whose query returned relevant results natively is a different signal than a shopper whose query returned zero results and was recovered by a fallback layer. Reporting them as one "search revenue" number hides which part of the funnel is actually doing the work — see zero-result query recovery for why that failure mode gets tracked separately. 3. **Match on session and query, not just session.** Tying a purchase back to the specific query — not just "search was used somewhere in this session" — is what makes the number defensible when someone asks how it was calculated. 4. **Report a share, not just an absolute figure.** "$5,700 in recovered cart value" only means something next to "out of what total." Pair every absolute number with the percentage of revenue or sessions it represents. 5. **Track it over a period, not a single day.** A one-day snapshot can be skewed by a promotion or traffic spike. A multi-week or multi-month window shows whether the attribution share is a steady pattern or a one-time blip. 6. **State what's excluded.** Note explicitly if the figure excludes autocomplete-only interactions, excludes browse-then-search sessions, or excludes a specific product category — exclusions change the number as much as inclusions do.

Worked Examples From Published Case Studies

These are real attribution figures from live Semantix deployments, each with its methodology stated alongside the number — the pattern this framework is built on.

**Weinroute:** In one month, Semantix-powered search accounted for 12% of online revenue. The attribution rule there was checkout-level: completed purchases where the path ran through Semantix-powered search, tracked from December 2025 once checkout attribution went live. The headline is the share of revenue, not a raw visit count.

**Wine House:** 39% of searches returned zero results before recovery; Search Saver's recovered queries went on to generate 26.5% of revenue in that period, about $5,700 in estimated cart value in one month. The methodology there is explicit: a search counts as zero-result recovery when the resulting product click is tagged with the recovery source, and an add-to-cart is matched to that search by the same session and exact query text within a short time window. Cart value is estimated from current catalog prices, since cart events don't store the price paid.

**Lisa Leonard:** Around 10% of zero-result queries are recovered each month, contributing roughly +1% in overall revenue. This is a smaller, steadier figure than Wine House's — a reminder that attributed revenue share depends heavily on how much of a store's traffic was hitting zero results in the first place, not just on how well the recovery layer performs.

Notice what's consistent across all three: each states its own window, its own match rule, and what's excluded (catalog SKUs and per-item pricing, in all three). That's what makes the numbers usable rather than just impressive-sounding.

Common Pitfalls That Inflate the Number

Any-touch attribution reported as if it were last-touch. A session where a shopper searched once, browsed for ten minutes, and then bought a different product is a weaker attribution claim than one where the purchased product came directly from the search results.
No time window stated. Without a cutoff, a purchase three days after an unrelated search query can get folded into the same bucket as an immediate add-to-cart.
Mixing recovery revenue with total search revenue. A store that reports "$5,700 recovered by zero-result fallback" as part of "12% of revenue from search" without separating the two is double-crediting the same dollars to two different mechanisms.
Comparing raw dollar figures across stores of different sizes. $5,700 means something different to a $20K-a-month store than a $2M-a-month one — always pair an absolute figure with the percentage share.
Treating a single good week as the baseline. A promotion, a viral post, or a seasonal spike can distort a short window. Report over at least a month before calling a figure representative.

Metrics to Track Alongside Revenue Share

Search-attributed revenue is more useful read next to a few supporting metrics than in isolation:

Search conversion rate — the share of search sessions that end in a purchase, tracked separately from overall site conversion since search sessions already carry more explicit intent.
Search exit rate — how often a shopper leaves shortly after searching without clicking a result, a leading indicator that usually moves before revenue share does.
Click-through rate on search results — see click-through rate — since a low CTR on a query that still converts elsewhere in the session complicates a last-touch attribution rule.
Zero-result rate — run your own with the zero-result calculator — since a high rate caps how much recovery-driven revenue is even possible to attribute.

The Bottom Line

Search-attributed revenue is only as trustworthy as the rule behind it. Define the attribution event before looking at the data, separate full search from zero-result recovery, match on session and exact query rather than "search happened somewhere in this session," and report a share alongside any absolute number. The published examples above all follow that pattern — state the method, then the figure, not the other way around.

Want help setting up attribution tracking that holds up when someone asks how the number was calculated? Book a demo and we'll walk through it against your own analytics.

Frequently asked questions

What is search-attributed revenue?

It's the share of revenue, purchases, or cart activity credited to a shopper's search interaction rather than browsing or another discovery path. The exact definition depends on the attribution rule a store applies — there's no single universal method.

How do you calculate search-attributed revenue?

Define an attribution rule first — typically a session-and-query match between a search event and a resulting add-to-cart or purchase within a set time window. Apply that rule consistently, then report the resulting share alongside the rule itself so the figure can be checked.

Why do different stores report such different search-attribution percentages?

Because the underlying rule differs: last-touch vs. any-touch, the time window used, whether autocomplete and zero-result recovery are counted as "search," and whether the match requires the exact query or just any activity in the same session. Two stores with similar underlying behavior can report very different numbers.

Can zero-result recovery revenue be counted as search revenue?

It can, but it should be reported separately, not folded into a single "search revenue" figure. Recovery revenue reflects a fallback layer catching failed queries; full search revenue reflects queries that worked correctly from the start. Combining them hides which mechanism is actually driving the number.

Does a higher search-attributed revenue percentage always mean search is performing better?

Not by itself. A high percentage can also mean a store has a high zero-result rate that a recovery layer is compensating for, or that the attribution rule is looser (any-touch, wide time window) than another store's. Compare the rule, not just the headline number.

Is search-attributed revenue the same as search conversion rate?

No. Search conversion rate measures the share of search sessions that end in a purchase. Search-attributed revenue measures the share of total revenue tied back to search activity through an attribution rule. A store can have a high search conversion rate on a small share of overall traffic and still show a modest attributed-revenue percentage.