-1.png?width=1132&height=292&name=Asset%201@2x%20(12)-1.png "Odins.ai Logo")
New to marketing mix modeling?
Marketing mix modeling (MMM) is a statistical method that estimates how much each marketing channel adds to sales, using spend and sales over time rather than tracking individual people. It works across digital and offline channels and needs no cookies. If you want the basics first, start with our complete guide to marketing mix modeling. This page assumes them and goes into what is specific to software companies.
The short answer
Model new paying customers, with the delay from first touch through trial to payment. Set the budget where the next paying customer still pays back inside the horizon you accept, using the marginal cost per customer rather than the blended average. Measure TV, podcasts and other awareness channels on paying customers too, since much of their effect arrives through branded search.
The sections below cover the outcome choice, setting the budget from payback, how awareness channels and branded search are read, the sales team's place in the model, and how to test channels the data can't settle.
Sign-ups, trials or paying customers?
Sign-ups are plentiful, arrive within hours of the ad and are easy to count. Many never pay. A channel that brings cheap sign-ups from people who were curious but never meant to buy will look like the best channel you have, and a budget set on cost per sign-up keeps feeding it.
Paying customers are the outcome that matters, and they arrive later and in smaller numbers. The model is set up with the delay from first touch to payment, so each customer is credited to the marketing that started it. Where paying customers are too few per week to learn from, a qualified step in between, such as an activated trial, is the workable choice.
| Outcome | Volume | Delay | Fit for budget decisions |
|---|---|---|---|
| Sign-ups | High | Hours to days | Weak: many never pay |
| Activated trials | Medium | Days | Good when paying customers are few |
| New paying customers | Lower | Weeks | Best when volume allows |
| New recurring revenue | Lower | Weeks | Best when plan sizes differ a lot |
Set the budget from payback
Software companies already think in payback: how many months of gross margin it takes for a customer to repay what it cost to win them. The same logic sets the budget level. Decide the payback horizon you accept, work out what a new customer brings in gross margin within it, and that is the most the next customer may cost.
The catch is which cost. Blended CAC, total spend divided by new customers, is an average. The next customer in a saturated channel can cost several times that. The model estimates the marginal cost per paying customer in each channel at today's spend, and the budget level is where it meets your ceiling. Below it, more spend pays back in time. Above it, it does not.
The blended CAC trap
A healthy blended CAC is often used as the argument for raising the budget. It may be right, but the blended number can't tell you. Raise the budget where the marginal cost is still below the ceiling, and hold it where it is not.
TV, podcasts and branded search
A software company that runs TV, radio or podcasts sees a familiar pattern: few people click anything, and branded search spikes. In click attribution, branded search gets the customer, and the awareness channel looks like it did nothing.
A marketing mix model reads it differently. When heavy TV lifts search sign-ups and paying customers in the following weeks, the model sees that TV went up and customers followed, and credits the lift to TV. Branded search then gets credit for what it adds on its own. Measured that way, awareness channels are judged on paying customers, which is the comparison that matters for the budget. Read how to measure radio and podcasts with MMM.
The sales team is part of the picture
Larger accounts often go through a sales team, and a strong quarter for sales outreach lifts results whatever the marketing did. Where sales activity moves results noticeably, it goes into the model as a control, so the marketing is not credited with the sales team's work.
We keep controls few, though. Every extra variable takes explanatory power away from the media, and anything used to explain the past has to be known when the model forecasts. Planned pricing changes and launches qualify. Smaller shifts are left to the trend.
Self-serve and larger accounts
Many software companies sell to both small businesses that sign up online and larger accounts that go through sales. The two respond to different channels on different timelines. Whether they need separate models depends on the data and the budget decision: separate when they are marketed separately and each has enough volume, together when one budget serves both.
Marketing mix modeling works best where there is a steady weekly flow of new customers, which suits self-serve and high-volume sales. Enterprise software with a handful of deals a month is better served by modeling qualified opportunities with a longer delay.
Test the channels the data can't settle
Some channels never vary enough on their own for the model to read them, typically ones that run at a steady level all year. A structured test fixes that: change the spend in one channel at a time, pulse it over a few weeks (some weeks normal, some off, some at double), and keep everything else stable. The model measures the result and gets sharper on that channel.
Pulsing suits channels that are easy to switch on and off, such as radio, podcasts and display. Auction channels optimized toward a target, like Performance Max, are tested by changing the target instead. Media platforms often co-fund a test when they know a marketing mix model is measuring it. See how we structure tests.
How Odins works for software companies
Odins is a managed marketing mix modeling platform. We connect the ad platforms and your customer data, build and maintain the model, and review every recommendation before it reaches you. Data flows in continuously and the model is retrained once a month, while scenarios and forecasts are yours to run whenever you need them, in the platform or from your own AI assistant.
Moving budget toward channels with headroom typically gives 5 to 15% more effect from the budget you already have. Companies like Tripletex and Adminkit run their marketing budgets on Odins. See the managed service.