As we continue into an era of rising media costs due to increased competition and fragmentation across audiences and touchpoints, marketers can no longer solely rely on traditional metrics such as return on ad spend (ROAS) to determine success and guide future investment decisions.
These traditional metrics highlight gaps in the full view of your marketing investment story. Coincidentally, these gaps are often brought about by traditional and expensive channels such as linear TV and print, making the true value of your investment difficult to track without a connection to digital engagement. But here are five ways to better understand your marketing performance.
-
Look Beyond ROAS for a Full-Funnel View
Let’s say you’re running a digital campaign with paid search. With search being a heavily intent-driven channel, the ROAS tends to run high due to a direct link to the post-click behavior. But with traditional ROAS, you have no additional view to what could have been the search driver. Is there TV in-market? Display? Billboards? Something awareness-focused? What about offline conversion activity?
Granted, many of the platform partners have solutions to try to overcome the multi-touch problem. These solutions are excellent for directional details when you’re serving ads in one consistent tech stack and have tagging in place from the view through to the site behavior. These data points tend to show aggregate, most-popular consumer paths and not much in terms of how investment drives that path.
Pro Tip: Don’t rely solely on ROAS from a single channel. Implement multi-touch attribution where possible to better tell the story of your marketing investments.
-
Solve Attribution Gaps With Identity and Data Partners
Log-level data provides a better view but is limited to trackable customer IDs and online-only behavior—not to mention limitations with users engaging across multiple devices or different locations, making identifying a single user’s experience hard to pin down.
Identity partners can help with matching IDs to activity. With the right resources and data, advertisers can tie online and offline conversions to users through an identity resolution partner and log-level ad serving data. This is an excellent sandbox to look at attribution and audiences, but the cost and data needs make this prohibitive for many small-to-mid-sized advertisers. In addition to being expensive, as security and privacy continue to take center stage, the fidelity of these IDs and the ability to match against ID partners is rapidly declining.
Pro Tip: If investing in identity resolution, prioritize partners that can help bridge online and offline behaviors, but prepare for a future with less reliance on individual tracking as privacy regulations continue to evolve.
-
Use Media Mix Modeling for Big-Picture Insights
Media mix modeling (MMM) has been available to advertisers for years. But with the advancements in AI and open-source solutions like Meta’s Robyn or Google Meridian, the ability to use MMM to inform your opportunity is more accessible and requires significantly less time with much lower costs. What helps push MMM to the next level is you don’t need attributed conversions to see the results. This means your online and offline investment decisions can be modeled to show how your overall activity ebbs and flows with the spend.
At the risk of over-simplifying the “math” that goes into creating an MMM, the easiest way to think about how to use the outputs is to identify your opportunity. MMM is not an attribution solution or something you should be looking at for day-to-day optimization but a tool to help you see how the variation in daily activity relates to investment.
Since MMM thrives on extensive historical data, the curves provide a view into the point of diminishing returns, setting your target investment at or below that point. These same curves show what your opportunity is to invest more or shift dollars between tactics.
Pro Tip: When implementing MMM, focus on collecting at least 18-24 months of historical data across all channels (including offline) to build robust models that accurately identify diminishing returns and highlight new opportunities.
-
Model Scenarios to Plan for the Future
Once you have the views into your opportunity, now what? Do you just take that on face value and continue the baseline investment? Using this data, the next step should be forecasting based on those witnessed curves. This forecasting exercise is called scenario modeling.
Scenario modeling takes the MMM outputs and lets you ask “what happens if …” If you move budget from social to CTV, for instance, you learn what you can expect. This shift from MMM looking backward and scenario modeling looking forward opens up a view of your total effectiveness and business outcomes. This provides a way to set expectations about what investments should drive for the business, without having to spend an additional dollar in-market. Once aligned and live, monitoring performance vs. expectation can provide a much cleaner view into areas of optimization or budget shifts.
Pro Tip: Run multiple scenarios before finalizing your media plan, testing different budget allocations across channels to identify the optimal mix that maximizes overall business outcomes rather than siloed channel metrics.
-
Validate Results With Incremental Testing
MMM and scenario modeling are great when looking back so you can plan for the future. But how do you validate these investment decisions are driving conversions that wouldn’t have been captured otherwise? This is where you start looking at incrementality.
Incremental lift tests can be run in-market, within a confined environment, before launching to the full market. This allows you to ensure results are reflective of the forecasts and gives you a chance to optimize or identify areas of opportunity ahead of launch. Incrementality shouldn’t just be a way to check your homework. Incremental measurement as an overall solution to attribution issues around reduced ID fidelity. Pulsing a channel off and on helps to solidify the contribution as increases in conversion activity changes.
Pro Tip: Design incremental lift tests strategically by selecting test and control groups that isolate the specific channel or tactic you’re evaluating. Consider geo-testing for channels that don’t easily allow for A/B testing on platforms.
Combine These Measurement Approaches for Maximum Impact
Having ROAS (or any single metric) as a signal of effectiveness is limiting. While there are many available options, the path forward is a combination of MMM, scenario modeling and incremental testing. Create a combined view of your dataset with clear and distinct usage for each.
MMM is great for identifying how channels are working together and where you may need to pull back. Scenario modeling lets you simulate your spend to set expectations and provide for balancing a variety of scenarios to best drive your business. And incrementality should be used whenever possible to validate and value your media investment in specific channels to your overall goals.
Build a measurement framework that incorporates all three approaches: Use MMM for strategic planning, scenario modeling for tactical budget allocation and incremental testing for ongoing validation, creating a virtuous cycle of continuous improvement in your measurement approach.
Scott Blessman is the VP of Analytics and Data Insights at Goodway Group.
Network
