Business Outcomes - The only common currency of Media Investments
In our Algorithmic Media Planning Primer, we described how Elsy’s algorithms are designed to optimize Media Investments towards Business Outcomes.
As logical as this may sound, this is far from standard in the marketplace today: most Media Investments are still planned and optimized towards Media Metrics; such as GRPs, Reach or Clicks. These Metrics are not directly comparable across Media Vehicles and cannot be tied to business performance, resulting in disconnected optimizations and sub-optimal investment decisions.
In an Algorithmic Media Planning approach, all proposed Media Investments are evaluated against expected Business Outcomes and budgets are allocated to maximize overall campaign returns. This is critical to support Zero-Based Budgeting approaches.
Is this just for Performance Marketers?
Optimizing towards business outcomes does not imply optimizing exclusively towards short term performance metrics. We need the ability to optimize across the full consumer journey (‘full funnel optimization’), not just bottom-funnel transactional metrics.
We believe that today, all media is performance media - whether designed to drive brand perception or transactional behaviour. The ability to understand the potential trade-offs between these often-competing goals is critical to being able to truly optimize the entire funnel and resulting returns.
Of course, the Measurement Plan needs to be tailored to each individual Advertiser. Very often this requires the combination of multiple forms of measurement, which leads us to…
How does this relate to Attribution Measurements?
You can think of Attribution Measurement as a range of backward-looking measurement techniques, which are designed to estimate the Business Impact of prior Media and Marketing investments.
The world of Attribution has evolved tremendously over the past 10 years; and is now very fragmented across a range of measurement techniques covering different Media Vehicles, types of Consumer responses and industries.
Algorithmic Media Planning leverages a new class of algorithms, which perform forward-looking Predictive Attribution – in other words they predict the likely ROI of proposed future Media Investments and optimize campaigns accordingly.
Of course, Media Planning Algorithms leverage the output of prior Attribution Measurements in order to predict and optimize future Media Investments. But very importantly, they augment such prior measurements with media marketplace data and brand characteristics in order to overcome the limitations of prior measurements such as lack of granularity, latency, incomplete coverage, etc.
How can Media ROI be predicted?
This will be the topic of a future post, but there are several critical components to it:
Multiple Sources of Data: We need to combine a number of different data sources to ‘overcome’ the limitations of attribution measurements, and drive granular and predictive media optimizations. These include media marketplace information, norms and benchmarks, brand fundamentals etc.
Multiple Sources of Attribution: We need the ability to combine (subject to availability) multiple forms of measurement such as Marketing Mix Models, Multi-Touch Attribution, Brand or Sales Lift Studies, Test Market Results etc. No single measurement model is perfect or all-encompassing.
ROI Decomposition: Rather than trying to predict a single ROI number; we predict all of the ‘component parts’ of ROI such as Budget Split, CPMs, Impressions, Reach Frequency, Targeting Accuracy, Inventory Quality, Response Rates etc. Each component part has its own predictive model (which gets better over time), then gets re-assembled into a predicted ROI.