Businesses – especially DTC brands – have never had so many media channels by which to promote their products. Upper-funnel channels like YouTube, TV, Podcasts, and influencers have been great for generating awareness while lower-funnel channels like branded search, email, and retargeting have guided prospects to conversion. Marketers have long tried to measure the return on advertising spend (ROAS) for each of these channels. Specifically, marketers want to know how effective spend on specific channels has been at driving sales. One method they use to do this is through media mix modeling (MMM).
Media mix models are statistical models – often time-series regressions – that use aggregated historical data to estimate how marketing activities influence sales.
Media Mix Modeling vs. Marketing Mix Modeling: What’s the Difference?
While often also called “marketing mix modeling”, media mix modeling is not the same. Marketing mix models look at the impact of the entire 4Ps marketing mix:
- Product
- Price
- Promotion
- Place (distribution)
Media mix modeling is a subset of marketing mix modeling, focusing specifically on advertising channels like TV, paid search, and display ads.
In modern digital marketing discussions, “MMM” typically refers to media mix modeling, even though the term originally meant marketing mix modeling.
We discussed marketing mix models in a previous post. In today’s post, we will discuss media mix modeling and how your DTC brand can benefit from it.
Why Attribution Alone Isn’t Enough
Marketers have traditionally used attribution models like first-touch, last-touch, and multi-touch to assign credit for conversion and direct marketing dollars, but that hasn’t been totally effective.
Attribution tends to overvalue lower-funnel channels and undervalue upper-funnel channels, so total reliance upon it will cause marketers to misallocate their media dollars. In addition, privacy changes like cookie deprecation and OS tracking restrictions are weakening their effectiveness.
Attribution models are great at answering questions like:
- Which channel deserves credit for this conversion?
- Did the creative for this campaign work?
But DTC marketers need different questions answered:
- How should I allocate next quarter’s marketing budget?
- Which channels have the greatest potential to maximize ROAS?
- Will spending $10,000 more on influencers generate far more sales?
DTC brands need to complement their attribution findings with insights that inform them of the incremental lift each additional dollar spent on a channel contributes to sales, and alert them to when they are spending too much on that channel.
The key distinction between attribution and MMM is that attribution tracks user-level behavior so that it can assign credit to the channel that drove conversion. MMM, on the other hand, uses aggregated data to measure incremental contribution, which is why it is so popular in an increasingly privacy-restricted environment.
How Media Mix Models Measure Incremental Impact
Attribution often tracks individual users; but media mix modeling analyzes aggregate time-series data to measure the incremental impact of each dollar spent on a specific channel.
Some inputs marketers might use in their model include:
- Weekly marketing spend by channel
- Price changes
- Promotions/discounting
- Sales or conversions
- Competitor activity (e.g., pricing, promotion, product launches)
- Adstock (time decay)
- Saturation (reach)
- Seasonality/holidays
- Economic conditions (e.g., unemployment, inflation, consumer confidence)
Media mix models use these inputs to measure:
- Incremental lift from each channel
- Diminishing returns (saturation)
- Lag effects (adstock)
We’ll address diminishing returns and lag effects later in this post; let’s now focus on incremental results.
Let’s assume a media mix model for your brand provides the following results:
Immediately, you notice that Influencers have the highest ROAS: on average, every dollar spent on influencer marketing generates about $4.10 in incremental sales.
But this does not mean you should put all your marketing dollars into influencers. After all, they account for only 7% of incremental sales.
This is because each channel has different audiences and not all your target prospects are equally represented within each. Moreover, the audiences of one channel may have different behavioral traits, so their responses and spending habits will vary from channel to channel.
However, the model can suggest how much of your media budget should be allocated to each channel to maximize overall ROAS.
This is what allows MMM to guide budget decisions.
How Media Mix Models Identify Diminishing Returns
Let’s assume you start spending money on a new channel to promote your DTC brand. Initial sales and conversions from that channel look encouraging, so you decide to spend more on that channel. You notice sales shoot up even more. You allocate another wave of dollars to that channel. Sales increase again, but not as much as during your last wave.
At some point, additional spending produces smaller marginal returns. When you first began promoting your brand within the new channel, you got exposure to some of the channel’s audience. When you spend more, you reach more people in that audience. But as you reach more people through the channel, the number of additional people you can reach goes down, so your message is not as effective. This is called saturation.
Many marketing channels follow saturation curves like this one:
As you can see, as channel spending increases from $50K to $100K, sales increase from $200K to $320K, so spending an additional $50K on the channel resulted in an additional $120K in incremental sales.
However, when spending increases from $100K to $200K, the impact is negligible: the additional $100K investment results in only $100K of incremental sales (from $320K to $420K).
Using response curves, MMM quantifies saturation so marketers can identify:
- Optimal spend levels
- Overspending thresholds
- Underfunded channels
The ability to pinpoint the cutoff at which additional spending on a channel provides no more benefit is another key strength of MMM.
Adstock: Why Advertising Effects Last Longer Than You Think
When exposed to a new marketing message, most consumers don’t immediately forget it. Depending on how engaging the message is and how engaged the audience is within the channel, consumers may remember it for several days or weeks. The message gradually fades from memory when no more exposure occurs.
Advertising effects that persist and decay over time are called adstock.
Adstock models measure carryover effects from previous periods. Media mix models contain adstock models to determine how long each channel’s advertising effects persist. Some channels have faster decay rates than others.
If a marketing mix model indicates that a channel has a high ROAS and low saturation, but a rapid adstock, it’s likely that the channel is extremely valuable but people forget the message fast. In such a channel, you would need to advertise more frequently.
How Media Mix Modeling Optimizes Your Marketing Budget
Marketing mix modeling then brings everything full circle by enabling you to run budget simulations based on channel responses, saturation, and adstock.
Let’s assume you have a $500,000 media budget. Without a model, you might choose to allocate your budget as follows:
MMM might suggest reallocating that same budget differently, as shown in this chart:
And your model might forecast a 12% incremental lift in revenues. Essentially the same budget, but more sales!
This is the benefit of MMM.
Why Media Mix Modeling Matters for DTC Brands
For DTC brands balancing acquisition, retention, and inventory planning, better allocation of your media budget can directly improve profitability. MMM helps marketers justify investing in brand-building channels that attribution often undervalues. And as cookies disappear and cross-device tracking weakens, MMM remains one of the most reliable ways to measure marketing impact.
Attribution tells you what happened. MMM tells you what to do next. To be a successful DTC marketer, you need both tools. MMM guides your future spending; smarter allocation drives your future growth.
