Pinterest is a growing visual search and shopping platform where people discover products, save ideas and plan purchases. For an ecommerce business or retailer, that creates an opportunity to reach customers while they are still deciding what to buy.
But how big is this opportunity exactly? How can marketers measure the true effectiveness of their investments into Pinterest advertising?
In this research article, we'll compare two alternative methods for measuring Pinterest advertising: Last-Click attribution, which attributes sales based on the last ad clicked before purchase, and Marketing Mix Modeling (MMM), which focuses on measuring the true incremental sales impact of advertising.
Across 18 advertisers in Europe and US, Marketing Mix Modeling showed that Pinterest's return on ad spend (iROAS) was 7.9x higher than Google Analytics 4's last-click ROAS, at the median. This comparison only includes online sales. Factoring in store purchases on top, would make the difference even higher.
If your business has physical stores, looking only at online sales misses a major part of your advertising impact. Because customers often discover products on Pinterest before buying them in-store, online-only reports overlook these offline purchases—making the performance gap even larger.
We compared MMM estimates with GA4's last-click reporting for 18 advertisers across Europe and US. The median ratio between MMM-reported incremental return on ad spend (iROAS) and last-click ROAS was 7.9x. This is significantly above the 4.2x multiplier for paid social we found as median in an earlier study.
This is also illustrated in the picture below.
What does this mean for advertisers? The median multiplier of 7.9x between Last-Click and Marketing Mix Modeling is high and shows that marketers interested in understanding the true revenue contribution of Pinterest should not rely on Last-Click, even in “getting a directional read”. You may cut campaigns whose contribution to Ecommerce sales is substantially larger than the attribution report suggests. Even worse, you might miss the opportunity to scale campaigns that are far from saturation.
When looking at individual advertisers, we saw high variance in the sample: the lowest multiplier was 2.1x and the highest 83x. High variance in the sample suggests that Last-Click’s ability to capture Pinterest's revenue contribution varies by business, and a general correction multiplier doesn't fix the problem. Your own multiplier will depend on your business and campaign mix.
1. Last-Click can see only part of the ecommerce conversions, typically missing ca. 20-25% of the ecommerce revenue. The difference comes from an increasing number of visitors declining analytics cookies. You can review the actual percentage for your business by comparing your Ecommerce sales and total revenue from the attribution tool.
2. Lack of incrementality: Last-Click assigns credit to the final click before purchase. Consider a shopper who discovers a product in a Pinterest ad, saves it, and later searches for the brand before buying. Last-click reporting may assign the sale to branded search even if the Pinterest ad helped create the demand.
MMM starts with the ecommerce sales recorded in your commerce platform or data warehouse, alongside Pinterest advertising data. The model analyzes how ecommerce sales change over time as Pinterest activity changes, accounting for other media and influences such as promotions and seasonality. It estimates the difference between sales with Pinterest advertising and the sales expected without it. That difference is Pinterest's estimated incremental ecommerce contribution.
MMM can also account for delayed responses to advertising. This matters for Pinterest, where a shopper may discover a product, save it, and return to buy later. Allowing for that delay helps the model estimate the sales effect across the period between advertising and purchase.
Relevant incrementality tests can help validate the estimates and inform MMM calibration. Together with the sales data, this provides a basis for estimating Pinterest's incremental revenue and iROAS for your business.
Let's take an example. You invest $100,000 in Pinterest campaigns and use the results to decide whether to continue. GA4 reports $50,000 of last-click revenue, giving Pinterest a ROAS of 0.5.
Based on that report, Pinterest appears to generate less revenue than the amount spent on advertising. Does this mean that Pinterest advertising should be stopped or scaled down? Here's the twist: the $50,000 is attributed revenue. It does not measure Pinterest's incremental contribution to sales.
Applying the study's 7.9x median ratio to this example implies that Pinterest was actually driving $395,000 in estimated incremental Ecommerce revenue. That corresponds to an iROAS of 3.95, or 4.0 when rounded.
An iROAS of 4.0 would give you a different basis for evaluating Pinterest than a last-click ROAS of 0.5.
If your business also operates physical stores, an Ecommerce-only analysis leaves out another part of advertising's potential sales impact. A shopper can discover a product on Pinterest and later buy it in a store. An online purchase report will miss that sale unless your measurement also includes offline outcomes.
To put that opportunity into context, we looked at the typical store vs. ecommerce sales impact of Paid Social that we see with Sellforte MMM implementations in retail:
The percentages explain how much incremental store sales an Ecommerce-only estimate can leave out. As another datapoint, a case study with C&A and Pinterest showed that more than 60% of Pinterests revenue impact was in offline stores.
Here's what the data means: for paid social performance campaigns, the ecommerce impact needs to be multiplied by 1.6x to arrive at full sales impact. For paid social awareness campaigns, the multiplier to reach full sales impact is 2.2x
Let's continue the earlier example, where $100,000 of Pinterest spend was driving $395,000 in incremental Ecommerce revenue. Assume, for illustration, that the Pinterest campaigns match the paid social awareness example in the previous section. Using the rounded 2.2x multiplier shown in the picture gives $869,000 in total incremental revenue, including store sales. That corresponds to an iROAS of 8.69, or 8.7 when rounded.
The example shows how the same investment can look very different as measurement expands: 0.5 ROAS in last-click reporting, around 4.0 iROAS for eCommerce, and around 8.7 iROAS when the assumed store-sales contribution is included.
To measure store sales, MMM uses revenue recorded in the retailer's point-of-sale systems as a separate sales outcome. Ecommerce and store revenue are modeled separately, providing an estimate of how Pinterest's impact splits between the sales channels.The model accounts for other drivers of store revenue, such as promotions, seasonal demand and weather, alongside the rest of the media mix.
Our Pinterest research found a substantial gap between GA4's last-click reporting and MMM's estimate of incremental Ecommerce sales:
By adopting MMM, marketers can fully harness Pinterest's potential to drive revenue growth. Book a demo with Sellforte to learn more.
Paul Arpikari is the Chief Commercial Officer at Sellforte, where he helps brands unlock the full value of their marketing. With a strong background in enterprise SaaS sales and a passion for marketing effectiveness, Paul builds and leads high-performing teams that turn strategy into results. He’s all about helping companies see marketing as an investment.