Today’s column is written by Sid Shah, Director of Business Analytics for Advertising Solutions at Adobe.
As a technical, data-savvy CMO, you want to make all key strategic decisions based on data and predictive analytics. To ensure this is possible, you’ve invested heavily in multiple solutions for measuring and optimizing marketing campaigns across all channels: you have a display-tracking solution, a search solution, a cross-channel measurement tag and a site-analytics platform, and you’ve even invested in an enterprise data cube to bring all your offline data together. But when you ask strategic questions – even basic ones – you still can’t get good answers.
Consider this simple example. A CMO wants to understand if his or her display campaigns are effective. The answer is often murky because display campaigns usually involve multiple vendors, each with its own unique tracking system. Adding further complexity, the search and display campaigns use different tracking systems. As a result, it’s impossible to build a cookie-level purchase path within display and across channels. Each tracking system doesn’t speak the language of the other. It is, as they say in Cantonese, like “a chicken talking to a duck.”
Meanwhile, CMOs face growing pressure to answer the media mix and attribution questions: What is the impact each channel has on my overall marketing performance? And how do I allocate budgets across channels, brands and geos to maximize all my business goals?
But when the basic tracking and data warehouses aren’t in place, it’s impossible to answer these questions. And if the data quality is poor, then all analysis based on that data is likely to be plain wrong.
It doesn’t have to be this way. It’s possible to get good data and to build a marketing-analytics engine that can effectively convert that data into the same language so decisions can be made based on all the available information.
What A Powerful Marketing-Analytics Engine Looks Like
The schematic below shows the basic building blocks of a good marketing-analytics engine. The foundation of the engine is the data-collection layer, in which data is collected from various marketing efforts as well as consumer responses at different points in the sales life cycle.
On the digital side, a unified cookie to capture the consumer journey within a channel – such as display – as well as across channels is essential. It’s also important to capture offline data to get the full picture; large multichannel retailers, especially, need to get offline store data to understand the impact that marketing has on their overall business. Finally, this first-party data should be augmented with data from third-party sources such as Nielsen to help put everything in context. And all this data needs to be captured and stored in a data warehouse.
On top of that, the analytics engine itself is made up of two key components: a cookie-transformation component that takes the cookie-level data and stiches it into the consumer journey, and an econometrics engine that marries offline data to the cookie-level information using pieces of information like time stamps. Econometrics becomes vital for measurement when there’s no direct way to assess the effect of an ad on a user, which is the case for offline marketing.
Once those data collection and analytics layers have been properly set up, the complex marketing questions can finally be answered. For instance:
- The cookie-level data can be combined with the econometrics to provide media mix forecasts.
- The cookie-level data itself can be used to understand the view-through attribution for display.
- The purchase path can be analyzed to understand the effect of advertising on customer lifetime value.
The possibilities are endless once the building blocks are in place. Other benefits of consolidating marketing data include:
1. Data as a common currency
One of the biggest advantages of a unified data warehouse or “data cube” is that it makes data the common and interchangeable language among different divisions and departments. This can help marketing address finance’s questions about whether the marketing team’s efforts are directly tied to conversions, for instance. With common cookie-level data and good algorithmic attribution, a unified data warehouse also eliminates the problem of “attribution wars” that often politicize marketing departments, in which each channel claims success over the others based on whichever attribution rule best suits them.
2. Closing the strategy execution gap
Another key problem marketers face is one of timing: Strategies are usually developed at the beginning of a fiscal year and executed later. And it might take weeks or months to assess whether a campaign was successful or not. However, by unifying the data onto one platform, you can understand the effects of a campaign quickly – and even correct its course mid-campaign by altering creative and adjusting budgets.
Big data presents a big opportunity for marketers but also a huge challenge. To become a truly data-driven marketing organization that can fully realize its revenue opportunities, it’s imperative to have consolidated, unified and standardized data on which to base your decisions. Otherwise, you’ll just be a chicken talking to a duck.