Retail media is showing explosive growth. WARC’s recent report 'Retail media’s path to consolidation' highlights that it is the fastest growing medium in advertising history, setting to top $128.2bn in 2023 and overtake linear TV by spend in the next few years.
This is despite the fact that 33% of advertisers claim they can’t spend more in retail media due to the lack of attribution and reporting capabilities, 56% cite 'the lack of standardisation of measurement and reporting' and 57% saying that retail media networks 'are not integrated by other tech' (Source IAB Europe).
The real issue is that retail media conversions are not being fed back to the originating source of the traffic, unless that is the retail media platform in question. To put it another way: Amazon is very good at attributing ads to Amazon Ads but is less good at attributing the sale if the users started their journey on a Google Shopping Ad or on Facebook/ Meta Ad. This matters because the full digital marketing ecosystem is not being measured properly and advertisers are spending marketing budgets on retail media somewhat blindly and taking the high ROI at face value.
This attribution piece is crucial for making data-driven decisions and optimising campaigns effectively but this can be challenging.
Complexities in attribution
Here’s what makes the attribution challenge complex.
Customer journeys toward a purchase aren’t necessarily linear or in the same session making it hard for the digital market tracking system to assign a ‘credit’ for the advertising spend to enable further optimisations. Customers might find a product through Facebook, learn more about it on Google, and then decide to buy it on Amazon. This complexity in the customer journey makes it difficult to determine which advertising efforts led to a sale as all the tech players in that customer’s journey will claim 100% of the impact rather than splitting the credit.
Attribution models are used to assign value to different touch points along the customer journey, as an alternative to ‘last-click wins’ which only values the very bottom of the marketing funnel. Attribution models range from first-click attribution (giving all credit to the first touchpoint) to ‘multi-touch’ attribution (sharing conversion credit between all touchpoints), with best possible optimisation obtained using Data-Driven Attribution which assign credit to touchpoints based on what actually works according to a statistical or machine learning method.
All such attribution models require the data to be passed between the platforms to correctly ascertain the amount of credit of each marketing £ spent. The challenge here is that the retail media platforms are not by default passing the conversion data back to the other platforms to enable out of the box attribution systems to work. Many advertisers are either manually stitching the data together in long winded manual processes, or worse, simply taking the high ROI’s of retail media at face value and investing way more in the channel than they’re worth.
Halo effects vs cannibalisation in the marketing mix
Brands spend money to drive sales, only to find themselves fighting a bidding war on their own brand terms and using social media to drive sales through wholesale and retail partners. The sales happen all the same, but the murky and complex interplay between direct and indirect retail channels means that money is definitely being wasted along the way. The whole point of improved attribution is optimising marketing spend to align with what is really driving sales through the whole funnel, but when the funnel leads customers down into a range of competing retail end points, out of the box Attribution tools just won’t give you the right answers,
Different margins
Brands selling through their own Direct to Consumer websites and channels will have a different margin for cost of goods than purchased through a retail media channel. However, many brands are simply not factoring in all the different costs of each channel at a margin level to make a true calculation on the advertising ROI.
How to address this challenge
Here at Croud we are pioneering new ways to answer this knotty challenge for our customers through custom attribution across a number of methodologies.
Custom Attribution
using the retail media API’s to manage sales conversion data to connect data together is the first step in this chain and it depends on which retail media channels are being used. If Amazon is a large part of your retail puzzle connecting to the Amazon API should be a priority to manage attribution across the funnel. Custom attribution is where your first party data is used to train a predictive data-driven attribution model that is unique to your brand, and allows you to split out more accurate marketing KPIs between different retailers and products.
Using custom attribution and market mix modelling together
If the retail channel you are using is a black box, with limited data visibility, then we build a base of attribution using accessible data sources and overlay an element of MMM modeling to manage external data such as brand impressions from retail media and social platforms where the customer journey is hidden by the retailer. This approach unifies the available data sources for precise short term decision making and crucially also helps answer the bigger long term questions needed to set marketing budgets for the year.
No two brands are quite alike, and the exact mix of direct and indirect retail, and the use of wholesalers and physical store networks, means that the widespread challenge of retail media attribution has many variants. A custom approach, with analytics built exclusively for you by an expert team of marketing analysts and data scientists, gives you the very best shot at cracking the retail media attribution code.
To learn more about marketing measurement, please get in touch with our Analytics team.
