Geoff Wolf, EVP Client Strategy
“Big Data” is a buzzword these days. More importantly, under the hood there have been new data emerging from so many sources, we are all scrambling to keep up. As new media channels have been born during the past decade, so have new databases that serve as repositories for all the new data. In some cases, each media channel has its own database. In other cases, the data is combined where it is easy to do so. In all cases, we need to figure out how to make sense of this escalating confusion of data.
The key problem is that each of these media channels wants to report on their sales with their own rules. In that scenario, it all adds up to more sales than what went into the bank account to begin with.
In a perfect world, all the data are combined in a single repository where all the metrics are normalized into a single “language,” so to speak, and represented relative to a return on your marketing investment.
The perfect solution may require more investment than you have available. In those cases, which represent the norm for most businesses, other solutions are used within a less complex database environment and even with spreadsheet tools. While not perfect, these solutions range from adequate to directional, and can be quite satisfactory for driving your marketing investment.
Regardless of the analytic processes you use, let’s look at some important foundations that must be established to maximize marketing investments.
Focus
What really matters is your definition of an acceptable ROI. Before turning your resources loose to work a bunch of data, think about what key performance indicators (KPIs) you need to focus on.
Define what matters most to your business so that your key performance metrics that result from all the analysis are focused. Your objectives must be defined to deliver the most value. The most common objectives are:
- Brand Awareness
- Top-Line Sales
- Profit
- Customer Acquisition
- Customer Retention
The reality is that focusing on a single priority yields more value for your business. If we aim for a blend of several objectives, the results will mirror that directive and success will be a blend of values across all the goals without any one of them achieving significant results.
Establish Rules
Once priorities are set, the next step is to establish rules around how the data will be defined and used.
The definitions have to make sense in the context of your business and the most important thing is to spend as much time as you need with your team to get these rules right. The rules are “right” when the team all agrees with them and they are tested.
Using someone else’s rules is not necessarily best, as each business has unique things happening. Start with industry best practice rules and then make them your own. It is possible that an element or two may have to be tweaked over time as rules are tested and results realized.
Let’s use allocation of email rules as an example. One option is to simply say that ALL sales arriving with a reference code embedded in an email will be allocated to the email channel. Let’s compare this with a second variation that says only the sales from those emails that arrive within the first 48 hours of the launch are allocated to email. The remaining sales arrived after another marketing touch occurred and are then allocated to the most recent mail piece (catalog or direct mail for example) per the order date represented by that clickthrough and subsequent sale.
The results vary by a significant amount in this example. In this case, perhaps you looked at the curve of email activity and found a clear front-loaded result. Most of the sales came in within 48 hours. So it makes sense that after 48 hours, other marketing pieces may have arrived before the email was opened and influenced the sale.
Unfortunately, a definitive best practice rule for allocating emails does not exist. Each business and its team must invest time into understanding what drives their brands and tweak industry rules accordingly.
Testing
Now that priorities and rules are set, drive the rules first in a virtual environment for a reality check. What happens if the rules are tweaked one way or the other? How does this affect the results, especially relative to what actions will be taken as a result of the analysis?
While this may seem like a lot of work and preparation before publishing an analysis, it serves to minimize the risk of making poor marketing decisions and finding out $100,000 later that there was a “rules malfunction” that distorted the reported results.
Finally, the rules process must include continual testing with real analysis to allow for both fine-tuning and changes in the marketplace. Setting rules is an ongoing process as the current economic and technological environment is changing rapidly around us all.
Tags: analytics, data, Geoff Wolf, multichannel