How to Use Marketing Mix Modeling

Marketing mix modeling (MMM) tools are a lot like professional knives—if you know what you’re doing, you’re going to get a lot done with them and the polish will be obvious in your final product. If you don’t, you may still accomplish something but they’re not going to help much with your efficiency. We know the power of the Prescient AI dashboard, so we’re intimately familiar with the campaign efficiency and revenue increase it can drive for our clients. But some of this does come down to knowing how to use the platform and what to do with an MMM generally.

Quick note that we’re building out resources for more of the logistical questions on how to use Prescient’s MMM. These guides, like how to use the dashboard to generate more revenue attributed to your paid media, can be found in our Learn section. In depth questions can also easily be covered with your Customer Success manager on a call to discuss the best options for your unique brand and apply them in real time.

What important information can MMMs provide?

MMMs like Prescient’s can look back (to tell you what most likely happened as a result of your paid media spend) and forward (to tell you what could happen if you adjust your spend or the campaigns that are running). More granularly, you’ll get important pieces of information from each of them.

When looking at attribution (what happened), you’ll get:

Revenue calculated by our MMM for each channel
Revenue calculated by our MMM for each campaign
Average ROAS calculated by our MMM for each channel
ROAS calculated by our MMM for each campaign
Halo or spillover effects calculated by our MMM for each campaign
- Revenue from organic traffic generated by each of your campaigns
- Revenue from branded search driven by each of your campaigns
- Revenue from direct search generated by each of your campaigns
- Revenue from your Amazon store driven by each of your campaigns

But our platform also aggregates reported data from each of your channels within the dashboard so you can look at these numbers against what the platforms are saying without navigating away from Prescient.

When looking at forecasts (what could happen), you’ll get:

Recommended spend amount for each campaign you want to consider adjusting
Predicted revenue generated by applying these budget changes overall
Predicted revenue generated by applying these budget changes by campaign

The recommended amount to spend on a campaign might be lower than you’re currently spending. Just as not spending enough on a campaign can lead to inefficient impressions and conversions, so too can spending beyond that campaign’s point of saturation.

What marketing mix modeling can do for brands

When a marketing mix model is built correctly for a brand’s media mix—we’ll go more into elastic vs non-elastic models and what fits your marketing data best in a future article—it can fix a lot of challenges that have long plagued marketers. Ultimately, marketing mix modeling provides more data to empower marketers and their teams to make strategic decisions to drive ROI. But in the day-to-day, that looks like a wide range of things, including:

Understanding how your marketing on other channels affects your Amazon performance
Figuring out how to maximize the returns on your paid media budget, even if you can’t increase your spend
Knowing when a campaign is truly saturated or needs a budget increase
Deciding where to cut budget for minimal loss or increased efficiency when needed
Feeling confident scaling up spend on promising campaigns
Understanding how an upcoming holiday or event could impact your ROAS and revenue
And, we’re just getting started here…

The best use of marketing mix modeling at a commerce brand looks at least close to this:

Baselining all of your performance and understanding discrepancies between what an MMM reports and what platforms report

Your paid media channels may be under- or over-reporting revenue gained from that platform. Most of these platforms rely on user tracking, which comes with increasing challenges, and—to put it plainly—they benefit from you thinking they’re more effective than they are so you’ll spend more with them and not another platform. Marketing mix modeling helps you understand right out of the gate whether your favorite channels are close to what probably happened mathematically, or whether you need to keep their numbers in context because they lean high or low.
Gut check what you’re seeing against your intuition as a marketer

Data you’re seeing in an MMM dashboard is only as good as the inputs and the models; there can even be issues with MMMs that boast high accuracy scores. That’s why it’s critical to gut check everything you’re seeing and flag anything that seems wrong to your customer success rep.
Taking time to understand what that means for your strategy assumptions

So, now you know that your Pinterest account is actually under-reporting your performance (just an example). Ideally, you pause to look at your marketing strategy and see if you should rethink your allocations with this new information. Maybe you’ve been holding back on YouTube because you thought it wasn’t efficient—but now you have new data that says otherwise.
Dive into the relationships between your campaigns, your channels, and your halo effects

With a good understanding of how your campaigns are performing according to the model and not the platform, it’s time to understand how they interact with one another, have ripple effects onto other channels (especially if you’re omnichannel and sell on Amazon or Walmart), and impact areas you might be ignoring like organic traffic, direct traffic, and branded search. The customer journey can be long and the most impactful interaction for your buyers doesn’t follow neat rules that other methods force on it, like last-click and MTA.
Identify optimization or growth goals for your brand that are realistic

Every brand is different: some want to work on optimizing one campaign first while others will choose to optimize several campaigns at once, which we call a scenario. Whatever the scope, we suggest thinking about a realistic goal you can hit over the next 14-28 days.
Get recommendations for shifting your budget allocation to move toward achieving your goals

This is exactly what our Optimizer is for. You’ll get recommendations for shifting or increasing your budget allocation depending on your inputs, and you’ll decide which changes you want to apply to your campaigns. We have a step-by-step guide for driving more revenue attributed to paid media here that walks you through using the Optimizer in this way.
Assess what happened and formulate your next move based on your findings

Check back to see how closely you stuck to the recommendations and how your outcomes compared to the predictions. From here, you can decide if you want to go through another round of working toward a modestly aggressive goal or really ramp things up. And if there’s ever a question of your next move, your Customer Success rep is always available to answer questions or help strategize.

Ideally, you and your team build trust in your MMM tool through optimizations like this so that you can eventually use the tool to help strategize larger decisions like how to allocate budget heading into a profit-heavy season like Black Friday/Cyber Monday. Even if a tool’s accuracy scores are high and it has attractive case studies, we know you want to build confidence first-hand, and this process is the way to do it.

Understanding ToFu and BoFu

Leveraging insights from your MMM can unlock a critical conversation in your marketing strategy: do we need to widen the top of the funnel because the bottom of the funnel has saturated? Your paid media strategy is probably always some combination of ToFu and BoFu campaigns—the ratios of each may change depending on the season or promotions you’re running, but they all come back to either acquiring new leads or converting people.

Most marketers have a great grasp on the interplay of these two campaign types, but may not fully understand the value of their top-of-funnel campaigns. That’s where an MMM like Prescient’s comes in. Our MMM reveals the impact of your ToFu campaigns on your organic and direct traffic, branded search, and Amazon traffic if you’re an omnichannel brand. You might skew your campaign type ratios if you see that your ToFu campaigns have positive ripple effects that you can now quantify.

Prescient’s MMM also offers saturation curves, which can show you whether it’s worth spending more on a campaign. These charts can reveal if your campaign is saturated and it’s time to spend less to optimize your efforts. They can also show if you need to increase spend to hit another increase in efficiency, potentially saving you from backing off or turning off a campaign that was in a trough and not fully saturated.

Getting more out of your MMM

We’re biased, but there’s a lot of exciting strategic work that can be done by leveraging an MMM for marketing growth. That’s why we understand when our clients get caught up in the daily or weekly opportunities in our dashboard. But MMMs can and should be used for more. As you’re ready to integrate it more into your organization, you’ll find more possibilities for using the tool and the data it provides to increase marketing efficiency. We love to see the Prescient dashboard used for tasks like:

Communicating more efficiently with leadership about marketing performance
Quantifying the bottom line value of your top of funnel marketing efforts
Understanding how your campaigns actually saturate and where you’re spending too much or too little
Making spending decisions to learn and build confidence strategically
Uncovering effective channels you may have written off in the past

You’ll find that with a robust MMM like Prescient’s, the more time you spend with your dashboard, the more you get out of it. MMMs unlock new, reliable data that helps marketers make better decisions with more confidence. Ultimately, and likely your biggest concern, they’re also essential tools for achieving some of your most important goals, like driving more revenue from your paid media spend.