We are currently living through a Cambrian-level explosion of marketing channels. It seems like there is a new channel or device popping up every year, each with a new set of metrics, media and strategies.
In the future, each marketing manager or analyst will be responsible for managing media across a greater number of different channels, networks and devices.
This calls for some changes in how we analyze and report on marketing results:
There will be a need to consolidate marketing metrics across different channels: hundreds of different metrics across dozens of channels is simply not manageable
Strategies will need to evolve from a channel-by-channel mindset to a cross-channel strategy that takes a holistic view of performance across media
Cross-channel analytics is oftentimes viewed as complex or convoluted, the domain of experts. This is not true – as with analytics for any single channel, it can be as simple or complex as you want it to be.
All it takes is a few additional steps to consolidate metrics in order to make sure you are comparing like vs. like. Below are a few steps that should help you get started.
1. Decide On Which Cross-Channel Metric(s) To Use
Without metrics to bridge the gap, there is no way to compare or aggregate performance across different channels. It is best if you can find a metric that represents the same action across channels (e.g., clicks mean pretty much the same thing regardless of media). If not, you may need to aggregate “similar” metrics into an uber-metric.
An example of an uber-metric is Avinash Kaushik’s idea of “amplification”: Facebook shares, Twitter retweets, and LinkedIn reshares are all engagement actions that spread content to a wider audience, so it makes sense to group them.
The criteria for a good-cross channel metric is that it fulfills the following:
Meaningful as a performance indicator
Occurs with some frequency across all media/channels
Not strongly biased for or against a media/channel over another
The last point deserves a bit more discussion, which takes us to the next step.
2. Be Aware Of Affinities For Each Channel On Different Metrics
If you use a conversion metric such as orders as your cross-channel metric of choice, media that is situated lower in the conversion funnel such as search will tend to have more affinity for tracked conversions compared to media higher up in the funnel, such as display or social.
There may also be more subtle biases for metrics and media that seem fair at first glance. For instance, tracked leads for paid vs. organic traffic may be biased if the lead form on the site is much more easily accessible via paid ads compared to internal links.
The “amplification” metric illustrates another example of a potential bias. Due to social norms and UI layout, likes tend to happen with much higher frequency than shares on Facebook, but the same does not hold true for favorites vs. retweets on Twitter.
All this means that it does not always make sense to directly compare metrics from different media, even if they represent similar actions.
3. Weigh Results Appropriately
This cross-channel bias affects your analysis when you are trying to make a comparison across media. In these cases, it is important to weigh the channels or metrics appropriately, to counterbalance the bias.
(Note that the bias does not affect your analysis when you are looking at the effect of one media on another – for example, the effect of a social activity spike on paid search sales. This is because the comparisons you make will usually be limited to time series within each media.)
In some cases, there are relatively simple ways to calculate an approximate weighting (to be addressed in a separate article). However, in many cases, the only way to obtain the “correct” weighting involves statistical or algorithmic analysis. You can get an overview of some of the possible methods here.
Unfortunately, these methods are oftentimes time-consuming to conduct, or painstaking to implement. The good news is that the nature of a particular metric or media does not change too much without a large change in reach, so the “correct” weighting, once computed, should stay more or less accurate for a good period of time.
After these three steps, you will have the following:
Metric(s) that are shared across channels or media
Appropriate weighting of metrics that allows you to compare results without bias
This should make cross-channel analysis much more manageable. Happy analyzing!
Comments