Explore the best creative works

Criteo: how Criteo and Groupon doubled down on data


Criteo agency won an award for 'Best use of Data' at the 2017 Drum Marketing Awards, for its work with Groupon. Here, they explain the strategy used to make this award winning idea.


Groupon is an e-commerce giant and has developed a global marketplace offering unique experiences, activities, things to do and discounts across mobile and web. Unlike typical ecommerce sites, the value of a particular offer to Groupon’s bottom line does not always correlate to the price of an offer. For example, a 50% off deal at a local eatery may be more profitable than a much higher-priced discount on a flat-screen TV.

What’s more, lower-cost local deals are more likely to attract repeat users who generate more lifetime value than a one-time electronics buyer might. Traditional data-driven bidding strategies aren’t designed for this kind of business model – focusing conversion and average order value data without factoring in margins or other unique KPIs the advertiser may prioritise.

Groupon wanted something designed for how it does business. It turned to trusted performance marketing partner for the answer. In less than a year, Criteo delivered a completely customised, fully automated pre-bid programmatic solution just for Groupon – one that used both user and product data to optimise for overall profitability, not just conversions.

With this custom-tailored programmatic solution in place, Groupon’s profitability soared by up to 30% quarter over quarter performance, and saw an increase of up to 10% Year on Year performance.


Groupon has been running Criteo Dynamic Retargeting campaigns since 2012 with excellent results. The Criteo performance marketing engine is known for significantly increasing sales and ROI by serving specific product recommendations, bids, and creatives that are most likely to result in a conversion based on Criteo's vast product, behavioural and intent data.

Groupon’s business model is unique. On its mobile and online marketplace, consumers can discover discount offers on everything from travel destinations to beauty experiences. The Criteo engine continuously captures and analyses data on more than 1.2bn active shoppers every month, including the products they view and purchase, the sites they visit and their site interactions, enabling Criteo to predict with great accuracy the optimal Groupon offers to recommend to each individual consumer.

But these offers vary enormously in how profitable they are for Groupon and margins don’t necessarily align with the price tag on the deal. What’s more, buyers of smaller, local offers are often more likely to return to the site again and again than, say, a one-time consumer who purchases an expensive cruise.

Groupon and Criteo had been addressing this issue by manually and reactively adjusting CPCs monthly based on shared data, in order to favour more profitable products or categories. Because this data was only applied post bid, the process limited Groupon’s ability to take immediate advantage of seasonal or market trends.


As a sophisticated marketing organisation, Groupon wanted a way to use both user and product data to automatically factor in the profitability of the deal to the overall optimisation strategy prior to the bid. The Criteo team immediately went to work.

Criteo developed a customised version of its optimisation engine to meet the specific programmatic needs of Groupon, using multiple data sets to optimise towards Groupon’s unique internal marketing KPIs.

This solution provided several very innovative capabilities:

● Margin-based optimisation. In addition to optimising toward each user’s propensity to buy and spend big, the engine factors in a value coefficient (from 1 to 10, 10 being most valuable) for each Groupon offer based on product-level profitability data.

● Pre-bid intelligence. The new optimisation engine operates pre-bid and in real time, eliminating manual data management and missed opportunities.

● Hands-off automation. Groupon shares product value data with Criteo which is used to generate the value coefficients used by the Criteo algorithm. No manual CPC adjustment or system maintenance is required.

● Micro-segmentation. The solution evaluates each bid based on the individual offer and user characteristics, making each bid hyper-granular with broader segmentation unnecessary.


In a clear A/B test that ran for six months across four markets (France, UK, Germany and Italy) starting in the summer of 2016, this innovative use of data delivered:

● Up to 5% higher profit margin and increase in Average cart value

● Up to 30% increase in new user activations year on year and quarter over quarter – truly new customer, first time converters

● Up to 20% increase in reactivation of lapsed users year on year and quarter over quarter overall resulting in up to a 10% increase to Groupon’s overall profitability. Due to the success of the A/B test, as of February 2017 this strategy has been applied to eight additional markets, including the United States.

By continuing to use The Drum, I accept the use of cookies as per The Drum's privacy policy