Why Is Marketing Attribution So Difficult?
Every business wants to understand where customers are coming from so they can optimise their marketing and advertising strategy. But behind every purchase is multiple website visits, devices and platforms, and tying together all of this data to get a clear picture of who is interested in what…can seem pretty impossible.
Today's Challenges
Consumers have greater choice than ever, sales cycle are getting longer (for the majority of categories) and new privacy laws make it more difficult to track people through their journey. Oh and not forgetting of course, that this still leaves you largely blind to offline customers. Different platforms use different metrics and it isn’t easy to compare them like-for-like. With a hazy picture of which marketing is actually working, many brands are blind to the true cost of acquisition.
Why It Might Be Happening
Ownership: Marketing may have had multiple interactions with a prospect before sales closes them, but knowing how to weigh that success between teams, campaigns & individuals is difficult…especially where people’s budgets/bonuses are at stake.
How You Attribute: Customers rarely convert from a single interaction and so that leaves businesses with a decision on how to attribute conversion. Let’s say for example:
- Emily clicks on your Facebook campaign on eco-friendly detergent…
- You don’t see her for the next 3 weeks…
- She then clicks on a link in an email…
- …and she makes her first purchase.
If you attributed based on last-touch then you would say that the email was what converted her, but if you attributed based on first-touch you would say the ebook. Both pieces of marketing have played an important role in converting the customer. You can imagine the impact that simple business rules would have on the outcome of your strategy, depending on how you attribute…
Let’s say that you measure the ROI of your ebook on the cost of production/marketing minus the sales generated within a six month period. What about all those people who convert after six months? Does this slower conversion mean that they’ve forgotten all about the ebook? Can you directly attribute that downloading the ebook led to the purchase? If your ebook is about laptops and they buy a desktop…does that still count? These numbers can massively influence the way you evaluate your marketing effectiveness and allocate funds, and are not even caused by bad data! Depending on how you attribute, when the cut-off point is and the channels you consider, it’s easy to come to different conclusions about the same conversion. Clear business rules need to be in place that are easy to identify.
External Factors: Most attribution models do not account for other external factors that influence sales and marketing effectiveness, such as pricing, promotions, seasonality, economy, and the wider world in general. It’s easy to overlook how big an impact these things can have to the effectiveness of campaigns.
Data exists in unintegrated silos: Customers don’t distinguish between channels, and their conversion paths regularly zig-zag. In line with this, data silos should be broken down, and channel data unified into a holistic, consistent format, to ensure that customer journeys can be tracked appropriately.
Infrastructure: Outdated infrastructure can be the achilles heel of attribution modelling. It’s important to have the right tools across data collection, integration, modelling, visualisation and forecasting. Data science and machine learning algorithms are rapidly advancing and can help to identify patterns and apply them onto your database.
What Could You Do Today?
As frustrating as it might be, 173tech would advise that there is no such thing as a perfect attribution. No matter how robust your data collection tools may be (and trust us, we can get pretty robust) you’ll never be able to capture every touchpoint for every customer. We typically recommend you start simple by adopting a last-click attribution model where you assume the final touchpoint is what ultimately influenced the lead’s decision to convert. While this may not give you the whole picture, it is data that is obtainable and easily understood. You can always estimate performance results where granular level data is not available.
The next step we would recommend is to deduplicate conversions with multiple touch-points to prevent under-reported costs.
How 173tech Could Help
As you expand your marketing offering it becomes more important to have visibility on the entire pipeline. We can create data pipelines that show you the full customer journey, test various attribution models to see which is the most appropriate to you, and apply complex algorithms to mitigate the impact of privacy policies. Get in touch with a friendly member of the team today to find out more!