There are various marketing channels today, some more suited for building awareness and brand equity while others are better for direct response. However, the reason marketers invest in any given marketing channel is the belief that it will eventually lead to revenue gains, whether in the short or long term.
In an ideal world, we would be able to assess the exact revenue impact of each marketing campaign regardless of channel; but, we know this to be difficult in reality. However, marketing has become increasingly quantified in the past decade, with digitization allowing for attribution of revenue to very granular marketing efforts.
The result is that now we live in a world filled with KPIs (key performance indicators) and proxy metrics — everything from high-funnel metrics such as impressions, to low-funnel metrics such as tracked revenue.
When surrounded by so many different metrics at various stages in the conversion funnel, how do we know when to use which metric, how to interpret them, and how to place them in the scope of actual marketing impact?
When & Why To Use Proxy Metrics
Though any proxy metric by itself only shows part of the larger picture, there are good reasons for using them as KPIs for assessing campaign performance, even when lower-funnel metrics are available. Four of the common reasons are below:
Short-Term Goals: There are cases where some proportion of the marketing budget is allocated not toward increasing revenue, but toward increasing higher-funnel metrics such as leads or registrations under the assumption that these efforts will eventually lead to a greater revenue base.
Trackability: For offline media, or when there is a long conversion funnel that involves non-digital steps such as call centers, lower-funnel metrics may become difficult to link back to marketing touchpoint(s). This will necessitate using higher-funnel metrics for assessing campaign performance.
Attribution Bias: Different marketing channels, or even individual campaigns within a marketing channel, are located at different points in the conversion funnel. As a result, assessing performance only using final conversion (revenue) results in attribution that is not reflective of actual impact.
Conversion Delay: If the conversion funnel is long, there may be a significant delay between top-of-funnel conversion to final conversion in the order of weeks or even months. For these cases, proxy metrics are useful for getting more timely feedback of campaign performance.
However, assessment of performance using proxy metrics must always been taken in the context of the fact that they are only a directional indicator of performance. They are usually correlated with final revenue, but do not represent the actual magnitude of impact.
Myths Around Marketing Metrics
Given the number of metrics floating around in marketing today, it is important to understand the limitations of what each metric can represent, and the caveats around interpretation. Below are some of the commonly believed myths that I’ve encountered.
Myth #1: Proxy Metrics Are Good Representations Of Marketing Impact
When looking at a given marketing campaign, we generally make the assumption that whatever metrics being used as KPIs are at least decent indicators of marketing impact. However, oftentimes this ends up not being the case, particularly with top-of-funnel metrics.
For example, one study found that click-through rate (CTR) in paid search, which is a commonly used digital marketing metric, only accounts for 4% of ROI (return on investment) change, where a 10% change in CTR only results in a 1% change in ROI.
Considering environmental changes and cross-effects from other marketing channels, it is difficult to assess whether an incremental change in performance for one marketing channel actually indicates the same directional impact on overall revenue, even for lower-funnel metrics.
Myth #2: Lower Funnel Metrics Are Always Better Than Higher Funnel Metrics
Since there are fewer conversions the lower you go in the funnel with potential converters being whittled down, it seems that a lower-funnel conversion may be more indicative of final revenue compared to a higher-funnel conversion. Indeed this is likely to be the case, but not guaranteed.
In the example is below, where two ads are compared in performance at three different steps of the funnel. While the total number of conversions shrink from 9 at the top to 3 at the bottom, the relative performance of Ad#1 to Ad#2 changes from 2:1 at the top, to 1:1 at the middle, going back to 2:1 at the bottom. This means that despite being higher in the funnel, leads turns out to be a better metric than applications to assess relative performance for closes.
This type of situation occurs with surprising frequency when the conversion funnel is finely sliced, due to differences in affinity of individual ads (or keywords, or targets) for driving higher-funnel vs. lower-funnel conversions.
Myth #3: Revenue Is The End-All-Be-All Of Metrics
When revenue is defined as total, non-attributed sales revenue, this may be a fair statement to make. However, revenue as typically represented in marketing is not total revenue, but tracked revenue for individual marketing channels, which does not differentiate between conversions that were impacted by marketing activity vs. those that just happened to touch it. Oftentimes, the revenue of individual channels does not even add up to the total revenue due to overlaps in attribution.
That being said, tracked revenue (with multi-touch attribution) is oftentimes the best practical metric to use for assessing marketing impact for digital media, due to the fact that going the final step in assessing marketing impact on total revenue will require extensive modeling efforts.
Generally speaking, as long as the proxy metric is chosen at as low a position in the conversion funnel as practically possible, a significant change in a channel-specific revenue is likely (though not guaranteed) to indicate a similar directional change in actual revenue impact.
Marketers Live In A World Of Proxy Metrics
Though we have come a long way in the past decade in being able to quantify marketing impact, what we have now is a mixed bag of good numbers and bad numbers. Some metrics are generally appropriate for use to represent marketing impact, and others should only be used in specialized contexts.
While some assumptions are necessary for practical and operational reasons, it is crucial to make careful decisions around choosing the best metric available for assessing campaign performance.
Comments