Why Your eCommerce ROAS is Declining and How to Fix It

ROAS was always the number that everyone paid attention to, until iOS 14 happened, then rising CPMs and then signal loss on Meta. Suddenly, due to this, the returns that once justified the spend started to decline. If you’re dealing with an eCommerce ROAS that’s declining, you’re not alone, and the issue is much easier to fix than most agencies will admit.

Signal loss isn’t the whole story

If you’re wondering, “Why is my Meta ads ROAS dropping in 2026?”, it isn’t all down to signal loss. The iOS 14.5 privacy update completely removed the ability for Meta to track most iPhone user behaviour, which reduced the volume and accuracy of conversion data that the Meta algorithm used to find buyers. This means that campaigns that were self-optimising with clean data now run on incomplete signals and overestimate their own performance.

The impact of this was a whole lot worse for brands that were heavily reliant on bottom-of-funnel Meta retargeting. These types of campaigns seemed to have ROAS because the reported conversions stayed high. However, in reality, many of those conversions were just view-through attributions from people who were going to buy anyway.

How attribution is lying to you

Every ad platform reports ROAS using its own attribution model, which is important to know when looking at how to fix a declining eCommerce ad performance. Meta credits itself with any purchase by someone who saw an ad in the last 7 days, while Google credits itself with any purchase by someone who clicked on an ad in the last 30 days. This means that both are simultaneously claiming the same purchase. When these are added up, you get a blended ROAS that looks good. But, when you look at your Shopify revenue, the number doesn’t match up.

The solution to this is quite simple. You need to not trust the platform numbers, but rather build an easy blended ROAS model using Shopify as the only source for your numbers. This means you should pay attention to the total Shopify revenue divided by the total ad spend. Doing this will give you a number that isn’t as attractive, but it’s a lot more accurate. Most brands that calculate it this way find out their actual ROAS is 20 to 40 percent lower than what their platforms tell them!

When we audit a Shopify account for the first time, the average gap between platform-reported ROAS and Shopify-calculated ROAS is 31 percent. That gap is revenue that no one is being held accountable for.

Fixing Meta through the structure

One of the main reasons brands experience declining Facebook ad performance is due to too many ad sets competing against each other for the same audience. This is a structural issue that leads to underperforming Meta campaigns. Meta’s algorithm tends to perform better with fewer and broader campaigns that are targeted to larger audiences. You need to combine ad sets, give the algorithm more room to find buyers and reduce the artificial audience fragmentation that can destroy your budget.

Using broad targeting on Meta in 2026 allows you to outperform detailed audience targeting in most tests for eCommerce brands with an existing purchase history. The algorithm knows a lot more about buyers than manually defined interest audiences. You need to create a new creative to test and a clear conversion event in GA4, while targeting a broader audience.

✓ Combine your campaigns. Have fewer ad sets, broader audiences and more budget per ad set.

✓ Use Meta's Advantage+ Shopping Campaigns for proven products and standard campaigns for testing.

✓ Upload your customer email list and Shopify order data as offline conversions monthly.

✓ Test creative systematically. Do this one variable at a time and have a minimum of 3,000 impressions before making a judgement.

We cross-reference your platform data with Shopify revenue and show you the real picture for free with no obligation.

Focus on Google Shopping feed quality before bidding strategy

When it comes to your Google Shopping ROAS falling, most of the issues you face are from the product feed, rather than the bidding strategy. Having a feed with missing product descriptions, wrong category mappings or images that don’t match Google’s specifications will restrain impression share before any bid setting makes a difference.

You need to audit your feed for title quality first. If you have a product title along the lines of “Blue Hoodie”, this won’t trigger specific purchase-intent searches. You need to redo these titles along the lines of “Men's Slim Fit Organic Cotton Hoodie in Navy Blue” because this will achieve purchase-intent searches. Your title structure should always follow Google’s recommended format, which is Brand + Product Type + Key Attributes.

The channel mix audit you should do before changing bids

If you’re wondering how to improve eCommerce ROAS, the best way to do this is to do a channel mix audit before changing bids. This means that before you even adjust any bids or budgets, you should build a simple channel contribution model. To do this, take your Shopify revenue for the last 90 days and attribute it by first-click source, last-click source and the channel each customer interacted with across their journey. When you do this, you typically find that email and organic search are generating revenue that the paid channels are claiming credit for.

Brands that discover this usually find that email is generating around 30 to 40 percent of revenue at minimal cost per pound, which means that they rarely cut their email investment. In fact, many aren’t investing in it at all because the ROAS isn’t visible on a platform dashboard!

Want to know what is actually causing your ROAS to fall?

We audit your Meta and Google accounts using Shopify as the source of truth. We tell you where the revenue gap is and what to fix first.