Hi, Happy Wednesday!
At the end of the day, media buying is an investment. Your goal is to turn $1 of ad spend into $1.20 of revenue or more.
Because it’s an investment, I tend to think of media buying a lot like building a balanced stock portfolio.
You want to identify and invest in underpriced positions (think of new marketing channels that have the potential to grow quickly) alongside placing more predictable bets (i.e. investing in the S&P 500 or various ETFs) to balance out your overall risk.
To me, Meta in this analogy is the S&P. It’s the tried and true king of DTC ad spend that delivers more predictable and reliable performance than other channels (especially at scale) but it’s also not going to deliver 10X returns in a short period of time.
On the other hand, smaller and newer channels like TikTok, podcast ads, influencers and affiliates and CTV are the growth stocks in your portfolio. They have more volatility, but if you invest at the right time, you can see huge returns.
The thing is, you can’t make good investment decisions without historical performance data and predictive models. In the stock market, you read earnings reports, look at fundamental analysis, technical analysis and sentiment related to new trends.
In marketing, you look at CAC, RoAS, MER and more across all of your marketing channels simultaneously modeled over time.
You study these metrics like a hawk to try to identify trends so that your CFO and CMO can produce forecasts of what they think will happen in the future and ultimately identify where you should either cut spend or double down.
You're asking yourself, is this marketing channel on a growth trajectory expected to scale, or are we at the peak now, and it’s time to sell this position and move on?
But the best investors don’t make these calls on a whim. They use tons of data and have dozens of signals/indicators to help them make these trades. The same level of analysis and trend prediction is now available to marketers in a big way.
Although the concept I’m about to share isn’t new, there have been some really powerful tools that apply this tech with the help of AI allowing brands better understand both channel specific performance alongside portfolio wide performance to help them build the most optimized ad spend portfolio possible.
The technology that’s powering all of this is called MMM.
An MMM stands for a Marketing Mix Modeling which is a statistical approach used to measure the impact of different marketing activities. It helps brands determine how much each channel (i.e., Facebook Ads, Google Ads, TV, influencer marketing, etc) contributes to revenue so they can optimize their budget allocation.
One tool that I’ve been using lately for this is called Prescient AI. It was built by a team of badass data scientists and engineers and it’s one of the most advanced MMMs I’ve seen yet. They’ve tracked over $1.5B in ad spend on their platform across 55+ data sources. Brands like Jones Road Beauty, Graza, Hexclad, and more are all using them.
How MMMs Work:
MMMs use historical data (often several years worth) and apply regression analysis to isolate the effect of different marketing efforts while controlling for external factors like seasonality, pricing changes, and trends.
Why It’s Important for Paid Spend Attribution:
Most digital attribution methods (like last-click or multi-touch attribution) rely on tracking user behavior through cookies or pixels which is less reliable due to:
- Privacy changes (iOS14, GDPR, CCPA)
- Walled gardens (Facebook & Google not sharing the full data)
- Cross-device & offline conversions (TV, in-store sales, etc are harder to track.)
-
Over-reporting by individual channels (i.e. individual ad platforms taking more credit than they should)
MMMs solve these issues by not requiring user-level tracking and instead analyzing aggregated data over time.
The Key Benefits of MMMs for Paid Spend:
- They measure incrementality – So it shows how much each channel truly contributes vs. what would have happened anyway.
-
They account for external factors – So it adjusts for seasonality, competitor actions, and economic shifts.
- They work without user tracking – So it’s a privacy-compliant alternative to pixel-based attribution.
- They help with overall budget allocation across channels – Since it helps identify which channels drive the best ROAS at different spend levels across the board.
Here’s an example:
Imagine a brand spends $500K on TikTok Ads, $300K on Google Ads, and $200K on TV ads in Q1.
After this, sales increased from $5M to $7M.
An MMM like Prescient determines that TikTok contributed $1M in topline revenue, Google $500K, TV $300K and the rest was organic growth or a halo effect that led to an uplift on another channel like Amazon or physical retail.
The brand can now adjust its budget accordingly, shifting spend from underperforming channels to the higher-ROAS ones.
Limitations of MMM:
That said, MMMs aren’t bullet proof. The drawback is that they typically require large amounts of historical data (usually 1-2 years) and they are less granular than pixel-based tracking (because it doesn’t track individual users). They also typically take weeks/months to set up vs. real-time attribution.
But Prescient, the tool I’ve been using, is different and unique in a few ways. Here are a couple things that stood out.
They give you:
- Same day reports. Most MMMs I’ve used have a lag when delivering insights. Prescient provides same day reports with their AI and expanded compute.
-
3 levels of insight. They provide channel, tactic, and campaign level analysis and reporting in their tool which gives you deeper insights than just total and aggregate lift.
-
The halo effect. You’ll also find reporting on your halo revenue, which comes as a result of impressions from people not clicking on a specific ad but seeing it in their feed, then coming back later directly, through branded or organic search, or to another channel like Amazon. In marketing, we call this the halo effect when sales on other platforms like Amazon or in store go up because of ads you are running on other channels creating that “halo” as you scale your paid spend.
- See all your channels, all in one place: You can easily monitor all of your marketing channels in one dashboard to review performance as needed by the day, week, or month. Having everything aggregated is incredibly helpful so you get the full picture within just one screen.
-
Actionable insights, quickly. It takes 15 minutes to sign up, connect your data sources, and get insights in a few days. Other MMMs take weeks to ramp up as they ingest your historical data. Prescient can be ready to use and deliver insights to you in just a few days.
In general, I’m a big fan of MMMs because they are currently one of the best tools in the market to measure true contribution margin across channels.
At the end of the day, if you are spending any kind of money on ads across channels, you need a tool that can help make spend recommendations in order to shift budgets to channels and campaigns that drive real revenue while reducing wasted spend on initiatives that don’t.
Prescient is that tool and can do this across 55+ data sources for your brand. This means paid social and search, TV, Amazon, native ads, and more.
I’ve been using it for some of the brands that we work with at Sharma Brands, and I’m loving it so far. So if you run ads on at least 3 channels, I think it’s definitely worth giving this a try. You can get a free demo here.
Anyway, that’s all for this week. I hope this quick deep dive on MMMs was helpful. It’s a pretty big topic in the world of marketing attribution right now but IMO, the MMM tech lives up to the hype.
When you have better data, you can make better decisions, and I think using tools like this can help you build a more profitable portfolio of marketing bets for your company this year. The goal is to eliminate wasted spend and only put dollars behind channels that actually drive real lift.
Alright, I’ll see you all back here for more insights on marketing and DTC later this week.
I hope you are out there crushing it!
See you soon.
-Nik
