How to use Google Analytics to analyze your search and browse audiences

This post was originally published on my substack, GTMOps.

One of the most surprising aspects of my conversations with customers since joining Algolia has been the number of teams who have no idea how much of their revenue is dependent on their product search functionality. Spoiler: it’s considerable.

The key to improving conversions at scale lies in driving users to search, personalized navigation, or recommendation experiences (rationale below), but how can you get a baseline for where you stand today? How many users search? What portion of users browse through category pages then buy? These two groups of users behave very differently. Telling them apart is the first step in isolating problems, opportunities, and insights.

By the end of this article, you’ll be putting together an analysis—from very standard GA reports—that looks a lot like this one:

Why do I need to know who’s searching?

(Or skip to next section to get the instructions)

Salesforce & Razorfish report that search engagement accounts on average for only?10% of users but 25% of the average eCommerce top line. The contribution is larger for B2B eComm (more customers know what they want), companies with large catalogues (more to find) and revenue over $30MM (scaled customer bases).

The key here is intent: eComm search users are 2-10x more likely to convert than users who are just browsing.

What about browse? The vast majority of most eCommerce customers are “browsing” only. Their conversion rates lag searchers, but users who navigate to Men’s> Clothing> Pants are performing product discovery as well, via filters. The same relevance considerations that are important for search are important for category-driven navigation experiences: the results presented to those users should optimize for conversions, incorporate business data, and personalize to the user’s history, just like search. They often don’t… and conversion suffers.

Both sets of users benefit from great product recommendations positioned in search, browse, or carousel experiences on Homepages, PLPs, and Cart pages. According to Salesforce Commerce, visits where the shopper clicked a recommendation comprise just 7% of all visits, but?24% of orders and 26% of revenue. Shoppers that use search and click a recommendation convert?3.7 times more often than those that only search,?and 4.2 times more on mobile. For more data take a look at?the SFDC report?I’m quoting (email wall).

The question is: between search, browse, and recommend, where should we focus as operators??The answer lies in benchmarking data.?In this post, we’ll use Google Analytics to understand the two major user groups:?search?and?browse?audiences. Once we’ve segmented them, we’ll pull together a KPI report that show us how they’re performing relative to each other, and compare those metrics to the market.

How to analyze your search and browse audiences in Google Analytics… in four steps.

1. Find the URL parameters or events that differentiate your search and browse experiences in the user journey.
2. Create two segments in your GA account, one for users whose sessions contain search (your searching audience), and one for users who’s sessions do not contain search (your browse audience). These audiences should be mutually exclusive and total to 100% of sessions.
3. Run your GA overview reports for the intended timeframe for Audience, Acquisition, Behavior, and eCommerce, and collect the data by segment.
4. Benchmark your performance based on your industry, and focus on the areas where conversion lags its potential.

We’ll be using Google’s?Analytics demo environment, built from the analysis of their?public-facing swag store, in combination with a public example from Algolia customer?Lacoste?in these steps. GA has multiple versions. Since most of us are still using Universal Analytics, and haven’t yet migrated to GA4, screenshots are from the universal analytics account. The same principals apply in GA4. Open the links and follow along!

Determining the parameters you will use to build your reports

The first step in building out your segments is to determine what?happens?on your site when a user performs a search. Do you throw an event? Does the URL string change? Are parameters added? Let’s look at two examples:

First up, Google Merchandise store. Pretend we’ve just performed a search for “shirt” and landed on this page. Notice anything? The URL has changed to include?“/asearch.html”.

Somebody should probably tell google this search isn’t doing what it’s supposed to do, by the way. Those don’t look like shirts.

This parameter will be present no matter what the search term is (unless it redirects). That means we can use it to identify the segment of users who have landed on this page, which is a mostly perfect proxy for anyone who completed a search! Note: the goal here isn’t perfect accuracy. We’re building heuristics for comparative performance of these audiences. You can get to an even finer level of detail by using events that occur every time a user performs a search instead, but that requires code.

Next, let’s take a look at Lacoste…

On the Lacoste site, the parameter that is added is “#query=”.

Your site may use different url components, but it’s likely that you’re using a similar structure. Again, you can work around the url strings if your site is throwing search events that capture every time a user performs a search.

Creating your segments

Now that we know what makes a search experience unique, we can use that to build out mutually exclusive search and browse segments. Click “add segment” from any page in google analytics, then click “new segment” and let’s get started:

Using the “conditions” tab, replicate the logic above. Insert the unique component of your search results URL?ONLY,?and make a mental note of the % of sessions represented by this group (in this case, 5.9%). We’ll use that number to check our totals.

Save that segment, then either copy it or create another. The only difference between this and the previous audience is that we’ll be changing “include” to “exclude.”

Our two segments (5.9% and 94.10% of sessions respectively) are mutually exclusive and total to 100%, so we can move on from here. If not, make sure the url component you’ve added is exactly the same for each, and make sure one is set to “exclude” and one “include.”

Note that the user counts will usually total to >100%, as we’ve used sessions, not users, to differentiate experiences.

Running your GA Overview reports

Now that our segments are done, the rest is easy. Click the “Conversions> Ecommerce” tab in the left hand sidebar and add the two segments you created to the report, as well as “all users”. Change the dates of the report to the time period you’d like to analyze.

First and most important: the eCommerce report.

The eCommerce report pulls the key revenue metrics for your business, Revenue, CVR, Transactions, and AOV.

What can we learn here? Even though search represents 6.27% of sessions, it pulls in 15% of google’s merchandise revenue. Transactions tell a similar story. For your organization, I’d expect AOV to be higher as well… usually by a factor of about 1.5X.

Next, we’ll take a look at the audience overview report:

The audience overview report contains details on users, sessions, pageviews, session duration, and bounce.

Google’s data here is illuminating of many of the broader trends we’d expect. Bounce rate falls by a factor of >6 when a user performs a search. Session duration goes up by >3x. Pages/session doubles.

Next, let’s take a look at acquisition. Navigate to Acquisition>Overview

The acquisition overview shows your conversion and bounce rates by web acquisition channel.

Here, we get a look at the comparative bounce and conversion rates by channel. Note how much higher the conversion rate is for the audience who search. Paid conversion is 3X higher. Looking for a way to make paid dollars more effective without raising budget? Get them to perform a search.

Finally: the behavior overview report.

The behavior overview gives you the metrics you need to understand what happens after a user lands on site. How long do they spend there? Bounce rate and exit rate are also reported here.

Here, we find details on TOP, Bounce, and Exit rate. Note: Everything is better when search is involved.

Pulling it all together

As a summary for your future to-do list, here are the steps above in list form:

1) On any GA Page, Add Segment > New Segment

• Insert the unique component of your search results URL?ONLY
• Save segment, copy or create another (change filter from?include?to?exclude?for the second)

2) Navigate to Conversions > Ecommerce

• Add the 2 segments you’ve created to the report and “all users”
• Change dates to the time period desired?
• Stats: revenue, conversion rate, AOV

3) Navigate to Audience > Overview

• Stats: Users, sessions, page views, unique users, etc

4) Navigate to Acquisition > Overview?

• Stats: Bounce and conversion by channel
• Bonus: check out how much more efficient your paid dollars are when they drive users to search!

5) Navigate to Behavior > Overview?

• Stats: time on page, bounce, exit rate

You’ve identified your audiences, created your segments, and run your reports. To drive this analysis home, look to benchmarking.

We recommend pulling the relevant data into a simple format like this one for ease of communication. Once there, note the differences between your organization and the market.

• What portion of your sales come from the search audience? If it’s below 25%, you’re missing opportunities.
• Do users who search convert at least 2x as often as browse? If not, you likely have a relevance issue.
• Is browse converting below average? Industry standard for retail hovers between 2-6% depending on the industry.
• What is the AOV impact of a user performing a search? Are related or complimentary products being suggested? Are results personalized?
• Is bounce rate too high for searching users? Typo tolerance or relevance issues are often the main culprits.

Have questions for us? Reach out!