Jan Bartczak, head of eCommerce at wallpaper and paint manufacturer and retailer, Graham & Brown, discusses the process of selecting a revenue attribution tool, and the value it has brought to the business.

“Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” While this is one of those “classical” marketing quotes – attributed to John Wannamaker, one of the pioneers of marketing in the US at the turn of the 19/20th century – my goodness, does it not just sound eerily on point in the midst of the digital age?

The rise of the internet and eCommerce has given us more and more touchpoints and channel data… and seemingly less and less understanding of spend efficiency. The Google search engine revolution, then paid search development, the rise of social media, Amazon turning into destroyer-of-worlds scale, the rise of mobile, paid social in full swing – while all these events gave consumers new ways to interact with products, brands (and other people), they have not always made it easier for the consumer to find what they want as quickly as possible.

One thing is for sure, these seismic changes have definitely not made things easier for a digital marketer. The opposite actually – today’s digital professionals are in the business of decoding more and more complex multichannel journeys, against a backdrop of the main channel owners keeping more and more of their data inside walled gardens.

The challenge
I have lived this complexity in my previous professional engagements, and the accurate measurement and improvement of ROI on marketing spend has been a focal point of my tenure at Graham & Brown in my capacity as head of eCommerce, where I am responsible for all aspects of digital and eCommerce. I wanted to be sure we invest in spend that allows our customers to find the right product for them, as quickly and as easily as possible. This would mean customers converting faster, improving our spend efficiency, while enabling us to serve those customers better!

Arguably, that stands at odds with the incentives of channel owners, who can be motivated by increased spend through their channel, and the yields it generates for them. We had a real customer centric business challenge, and revenue attribution became a tool we are using to solve it.

Just to set the stage for our challenge, Graham & Brown is a wallpaper and paint manufacturer and retailer based in Lancashire. We have our own fantastic design studio, full of true artists creating striking pieces of wall coverings, and we proudly manufacture our product in the UK.

While Graham & Brown has long been a supplier of choice for DIY retailers (with our Superfresco and Superfresco Easy brands), we also have a direct-to-consumer channel, centered around our premium Graham & Brown brand product range.

We craft our product with passion to help our customers create their perfect loving home. Wallcoverings are often the focal point of a room, at the very least giving it its character and feel. As such, it is a considered, long, multi-touch (and often multichannel) journey that leads to purchase.

We faced a challenge in the business on how we allocate our finite resources between brand-building and performance marketing activity. There have been several questions about the effectiveness and efficiency of different types of activity and channels. It very clearly became obvious that we needed tools that will give us the insight needed to make the right investments. In order to achieve this, we have invested in a data-driven, multichannel attribution tool.

Starting the journey
Our journey into understanding the drivers of our revenue began months and months before we got to an attribution tool, and it was a journey along two paths. The first one was a review of what tools and intelligence were available, before committing to a custom revenue attribution tool. The second was developing and driving the strategy of how we want to manage our digital marketing and its related data.

While advertisers and retailers have more and more tools at their disposal, for my team and myself, they turned out to be thoroughly insufficient in answering any of the major questions before us. Last click attribution, which might have some utility for short path, distress purchases, completely fails at accounting for the value of different clicks in the path. The other attribution methods offered as default by Google are position-based in Analytics (such as first click, linear, time decay); or data-driven in Ads.

The former fail on assigning value to given clicks without considering user intent or any channel interaction. The latter only take into account Google Ads data, optimising only a piece of spend/channel pie. Once you move out of channel, looking at social/paid social it does not get any better, with Facebook attributing value to its ads post-impression and post-click in a fairly aggressive way.

A similar approach is shared by DSPs and affiliate networks, with full value attributed post-impression and post-click for a given look-up window. As an advertiser, we then ended up with multiple channels taking claim for a sale, plus last click values. It became clear we needed a more sophisticated tool. Linear regression has been widely used to measure the value of advertising, so it was an option. However, as that methodology does not take into account cooperation/interaction between the channels, we decided it was not right for us. We decided to invest in a multi-touch, algorithmic, data-driven attribution model.

In conjunction with that journey, we have developed and executed the strategy of how we are going to manage our digital marketing. While I have a preference for building competency in house to manage digital marketing directly, this has not been about being dogmatic about in house versus agency. The main premise of this exercise was to ensure all our data is properly organised, tagged up and clearly classified.

Building an in-house team also allowed for having resource that is able to interpret and leverage the information outputted from the attribution tool. Businesses that work more extensively with agencies need to ensure they have this data consistency, as well as, having partners that will support them in disseminating the insights, rather than a narrow channel managing view that has been a stereotypical bane of the client/agency relationships.

Questions answered… more questions asked
It is fair to say that it took myself, my team and our business a while to get to a position where we felt ready to make that leap. Once we did though, things went fast. After a number of enquires, we identified a few promising providers and after a fast but intense process, we started working with RO EYE, adopting their award-winning SaaS multi-touch attribution platform, Single View.  singleview.media. Without divulging any sensitive information, this tool uses a proprietary algorithm which is based on the principles of cooperative game theory principles derived by L Shapley, such as Coalition Game.

RO EYE impressed us not only with power of the tool to assign value across complex paths, it also takes channel synergies properly into account, while being very intuitive and user-friendly – a truly winning combination! I also felt we were well aligned culturally as partners, and I really appreciated the level of care and support we got from day one up to now.

Proper revenue attribution is not for the faint hearted, in the sense that it takes a few fairly unnerving months before any meaningful results are available, as the algorithm is in “learning mode”. The advantage was that Single View does the learning on each client’s data specifically, which meant the results which would accurately reflect our customers journey.

Once the learning phase was over, it allowed us to start to build a much fuller picture of our activity. We were able to understand the value of our paid-for activity. As an example, we were able to assess the value of our affiliate channel, which, while completely different from last click, strongly pointed to incrementality of its activity. That was obviously based on our data. Affiliates remain a controversial channel. For a broad, marketwide look I would recommend this comprehensive white paper from RO EYE. 

The tool gave us a granular understanding of our paid search activity, highlighting which campaigns contribute strongly and at different points of the funnel. Having robust revenue attribution also allows for measuring and assigning value to channels whose prominence is mainly further up the funnel, such as certain social platforms and digital brand-build activity.

The journey has not been without challenges though. Wallpaper and paint are products with a long and very complex, non-linear journey. The length of the consideration period requires taking a bit of a step back to ensure that one has a meaningful timeframe of data to make decisions. The fact that some customers buy wallpaper samples, paint tester pots or use a paint card before buying adds an extra layer of complexity. Additionally, it can take weeks or even months between an initial sample order and the subsequent selection of wall coverings. Sample orders, technically being low-value purchases, but which can relate to another future purchase, are certainly a challenge in terms of properly valuating them, which is not present in most other retail environments.

The journey, not the destination
While there are certainly challenges for our team at Graham & Brown, as well as for our attribution partner, for me personally that is something that makes me spring out of bed in the morning – the chance to figure out another riddle that will help us understand our customers better.

Having a robust attribution model in place allows for not only that, but also gives opportunities to test activating new channels and new platforms, allowing us to measure their actual value, beyond the rudimentary measurements available otherwise.

One word of caution though – revenue attribution, or any other insights tool, is not a silver bullet that will make business challenges instantly go away or allow for effortless attainment of business objectives. It is also not something that should replace strategy and thinking, but rather aid it – it is a great tool, but just that.

In the case of Graham & Brown, we are able to utilise that tool on multiple levels. The more tactical uses range from leveraging the observed cross-channel interactions, to exploiting the granularity of the data to act on specific ad groups, campaigns, publishers etc. On a strategic level, it brings us so much closer to understanding the value of top of funnel activity. It also allows us to map the shape and composition of our digital funnel, which in turn allows us an informed review of our broad business strategy and what spend/support is needed to achieve our objectives.

While some might be taken back by how this is an ever-flowing process than never ends, I for one am enjoying the journey immensely and I am looking forward to new answers… and a lot more new questions!