Most merchants treat Google Ads as a machine that buys sales. You put money in, orders come out, and the only question is whether the return on ad spend is acceptable. But your ads are buying something else at the same time, something most advertisers throw away: information about what customers wanted versus what they actually bought.
When you add cart data to your Google Ads conversions, you stop seeing just “this click produced a sale.” You see exactly which products ended up in the cart. And once you can see that, three patterns show up on almost every ecommerce store. Customers click one product and buy a different one alongside it. Customers click a cheap product and leave with an expensive one. And customers buy the product you advertised, but in a different color or size than the one shown in your ad.
Each of these patterns points to a specific, practical change you can make to your product landing pages. In this article I will walk through all three, with the exact changes I recommend and the reasoning behind them. If you prefer to watch, the video version below demonstrates every change on a live demo store.
First, a Quick Word on Where This Data Comes From
Google Ads does have a report for this: the “Cart items cross-sold” data available to advertisers who send cart data with their conversions. The catch is that most stores never send that data in the first place, so the report sits empty. In earlier an earlier course I covered how to set up cart data tracking and how to fix it when the report shows nothing, so I will not repeat that here. For this article, assume the data is flowing. The tracking solution and the visualization app I use throughout are both available at MerchantCodex, which turns the raw report into a clear list of cross-sells, upsells, and variant cross-sells in seconds for Shopify users.
Now to the part that actually makes you money: what to do with it.
Scenario 1: Cross-Sells, Customers Click Product A and Buy Product B
Alpine Flask 750ml
Complete the set — save 10%
Customers who landed here also bought
Deep-Clean Bottle Brush
Silicone Protective Boot
Straw Lid 2-Pack
Carabiner Handle Clip
This is the most common pattern you will find. Someone clicks your ad for a water bottle and the order that follows contains the bottle plus a cleaning brush. Or they click the bottle and buy only the brush. Either way, your cart data is telling you that product B has proven demand from product A’s traffic. Not assumed demand, not “customers who viewed this also viewed” demand generated by an algorithm guessing from category similarity. Actual purchase behavior, from your own paid traffic, with revenue attached.
The first change to make is a “frequently bought together” section on product A’s page, populated with the actual winners from your data. Most Shopify themes and recommendation apps fill these sections with lookalike products and hope for the best. Your cart data lets you fill them with products that have already earned their place with real orders. The difference in conversion between a guessed recommendation and a proven one is substantial, because the proven one matches what the customer was statistically likely to want anyway.
The second change is a bundle. If a meaningful percentage of buyers add the brush to the bottle order, create a bottle-plus-brush bundle with a small discount and place it near the add-to-cart button. You are not trying to convince anyone of anything here. Customers were already buying these together; the bundle simply removes steps and lifts your average order value in the process.
The third change is the smallest: mention the cross-sell product inside product A’s description where it fits naturally, with a link. One sentence like “pairs perfectly with our deep-clean brush” shortens a path the customer was going to walk anyway.
There is also a strategic insight hiding in this data that can save you from a costly mistake. Sometimes the data shows customers clicking product A but buying product B instead of it. In your standard Google Ads reports, product A looks terrible: plenty of clicks, few conversions, poor ROAS. The obvious move is to pause it. But the cart data reveals that product A is a door-opener. It wins the click and hands the customer to your bestseller. Pause it and your “good” product’s sales quietly drop with nothing in the reports to explain why. Without cart data, this mistake is almost impossible to avoid, because the report that would warn you does not exist.
Scenario 2: Upsells, Customers Click Cheap and Buy Expensive
Trail Tumbler — Basic
Just +$15 gets you the Pro
Double the insulation time, leakproof flip lid, lifetime warranty.
Find your Trail Tumbler
Basic
- Capacity 16 oz
- Cold retention 6 hrs
- Lid Press-fit
- Insulation Single wall
- Warranty 1 year
Pro
- Capacity 20 oz
- Cold retention 12 hrs
- Lid Leakproof flip
- Insulation Double wall vacuum
- Warranty Lifetime
Elite
- Capacity 24 oz
- Cold retention 24 hrs
- Lid Magnetic 2-in-1
- Insulation Copper-lined vacuum
- Warranty Lifetime + loss cover
Price wins clicks in Google Shopping. Your cheapest products attract a disproportionate share of traffic, and in a standard report they often look mediocre: low order values, thin margins, unimpressive ROAS. Cart data frequently tells the opposite story. A significant share of the people who click the entry-level product end up buying the premium version. Customers are upgrading themselves; your only job is to stop making them hunt for the upgrade.
Start with a callout near the price of the cheaper product. The framing matters more than you might expect: “just $15 more gets you the Pro” performs very differently from “the Pro costs $34,” even though the math is identical. Presenting the upgrade as the difference rather than the full price makes it feel roughly half the size.
Then add a comparison table further down the page, the classic good, better, best layout. Put the tiers side by side with the handful of attributes customers care about, and highlight the tier your data shows people actually choose. If 62 percent of buyers pick the Pro, label it “most popular” and mean it. Comparison tables work on two groups at once: customers who were already going to upgrade do it in ten seconds instead of five minutes of tab-switching, and a percentage of customers who were not planning to upgrade see the columns next to each other and decide the middle tier is obviously the sensible choice.
The strategy layer here mirrors scenario one. Keep advertising the cheap product. Its low price is what wins the auction and the click; that is its job. Stop thinking of it as a product and start thinking of it as a doorway, then design the page behind the doorway to sell the range. That nineteen-dollar item with disappointing ROAS may be the best salesperson on your store, and cart data is the only report that will ever show you that.
Scenario 3: Variant Cross-Sells, Right Product, Wrong Color
Summit Bottle 1L
This is the cheapest fix of the three, and often the fastest to pay off. Your ad shows the blue bottle. Your feed contains the blue bottle. But when you look at the cart data, the majority of buyers choose black. Every click starts with a mismatch: the customer lands on a blue product page and has to work, find the color option, open a dropdown, and hope the color they want exists. Every second of that work is a chance to bounce, and you are paying for all of those seconds.
On the landing page, make the variant selector impossible to miss. Visual swatches above the fold, not a dropdown buried under the description. Put a “bestseller” badge on the variant your data proves people choose, and make sure the main product image changes when a variant is selected. That last point sounds obvious, yet a surprising number of stores show a single photo regardless of the selection.
The bigger fix is in your feed. In Shopping ads, the variant you submit determines the image shoppers see in the ad itself. If black is what people buy, black is what the ad should show. This is often a one-line change to the feed, and it improves click-through rate and conversion rate at the same time, because the ad, the landing page, and the customer’s intent finally match. Consider what the alternative means: you may be spending your entire budget showing customers a color they do not want, and no standard Google Ads report will ever tell you.
Two Wins From One Data Source
Every change described above comes from the same place: the gap between what customers clicked and what they bought. Reading that gap gives you two compounding wins. Your landing pages convert better on the traffic you are already paying for, and your bidding gets smarter because you finally see each product’s true value, door-openers included, instead of judging every product by its own last-click numbers.
None of this requires guesswork or generic best practices copied from a blog post. Your customers have already voted with their wallets. The work is simply rebuilding your pages around the results.
Getting the Data for Your Store
Everything in this article depends on cart data flowing into Google Ads, and on being able to read it quickly. MerchantCodex covers both: the tracking solution sends cart item data with your conversions, and the app classifies the results into cross-sells, upsells, and variant cross-sells automatically, so you spend your time making changes instead of wrestling with spreadsheets. If your “Cart items cross-sold” report is currently empty, that is the place to start, and the earlier parts of this series walk through the setup step by step.