A Simple Framework for Marketing Budget Allocation in 2026

The 70/20/10 rule says put 70% of your budget in proven channels, 20% in promising ones, and 10% in experiments. It has been repeated in every marketing textbook for two decades.

It is also useless.

Why Rules of Thumb Fail

The 70/20/10 rule assumes you know which channels are "proven." Based on what? Platform self-reported ROAS? Last-click attribution? Your gut feeling from 2023?

A channel that returned 4x two years ago might be returning 1.2x today because CPMs doubled and your audience saturated. Rules of thumb do not account for diminishing returns, seasonal shifts, or competitive changes. They give you a false sense of rigor.

In reality, the right allocation for your business depends on your specific saturation curves, your margin structure, your growth targets, and the current market conditions. No rule of thumb can capture that.

A Better Framework: Three Layers

Here is a practical approach that works whether you have an MMM tool or just a spreadsheet.

Layer 1: Set your floor spend. For each channel, what is the minimum spend needed to maintain presence and collect data? For most digital channels, this is the budget level below which the platform algorithm cannot optimize. On Google Ads, that might be $5K/month for Search. On Meta, $3K/month per campaign. Below these floors, you are wasting money because the platform cannot learn.

Layer 2: Allocate by marginal ROAS. For each channel, estimate the return on the next dollar spent (not the average return). If Google Ads averages 3x ROAS but the marginal ROAS at your current spend level is 1.1x, that channel is nearly saturated. If TikTok averages 2x but marginal ROAS is still 2x, there is room to grow. Put your next dollar where marginal ROAS is highest.

Layer 3: Reserve 10-15% for testing. This is the one part of 70/20/10 that still makes sense. Allocate a fixed testing budget to try new channels, new audiences, or new formats. Run each test for at least 6 weeks with enough budget to reach statistical significance.

The Spreadsheet Version

If you do not have an MMM tool yet, you can approximate this with a simple spreadsheet:

• Column A: Channel name
• Column B: Current monthly spend
• Column C: Platform-reported conversions (discount by 30-50% for overlap)
• Column D: Estimated revenue per conversion
• Column E: Rough ROAS (D*C/B)
• Column F: Is the channel still scaling? (Yes/No, based on whether increasing spend last quarter improved or hurt ROAS)

Channels with high ROAS and "Yes" in Column F are where you increase spend. Channels with declining ROAS and "No" are where you pull back. It is crude, but it beats gut instinct.

The MMM Version

MMM does all of the above automatically and more accurately. It gives you actual response curves per channel, so you know exactly where diminishing returns start. It accounts for adstock (the lagged effect of past spend). It adjusts for seasonality so you are not confusing Q4 holiday demand with your September campaign performance.

The output is a recommended allocation that maximizes total revenue for a given budget, or minimizes spend for a target revenue goal. No guessing, no discounting platform numbers by arbitrary percentages.

You can see what this looks like in practice in our sample MMM report, which includes a budget optimizer showing optimal allocation across channels.

Seasonal Adjustment: The Forgotten Step

Most teams set their budget in January and barely touch it until next January. This ignores one of the biggest drivers of marketing efficiency: seasonality.

CPMs on Meta drop 25-40% in January after the Q4 auction frenzy. Google Ads CPCs in B2B spike in September and January (budget planning season). TikTok CPMs are still 30-50% lower than Meta for many audiences.

A smart allocation framework adjusts monthly. Spend heavier when CPMs are low and your audience is receptive. Pull back when costs spike and efficiency drops. Even without an MMM, you can track CPM trends and shift 10-20% of budget toward cheaper periods.