Why Elevation Testing Alone Won’t Fix Your Paid Media Budget – The Missing Metric
Growth testing has become the default answer to most direct-to-consumer problems. The adjective platform does not agree with it; Meta and Google often both claim credit for the same conversion. Not to mention the studies we’ve done reviewing one transaction at a time to determine whether organic search or Google Shopping transactions were targeted.
Somewhere in that noise sits the question of actually allocating a paid media budget.
The standard pitch is that the climb cuts through it. Conduct lift research, determine which channels are generating demand versus yield, and reallocate funds accordingly. Most of the content you’ll find growing up in the last few years falls somewhere in that area. That framework isn’t perfect, and working on it has likely led growth-stage brands to bad decisions. The most common is to cut the top funnel channels that don’t pass the independent height test, only to look at the total revenue because those channels were doing work that no other channel could see.
Conversation needs a different anchor.
Why Nurturing Alone Doesn’t Answer the Allocation Question
Growth measures the causal impact of a particular channel or campaign. That’s real useful information, but it’s not the same as understanding how marketing contributes to the business as a whole.
Consider a customer who sees a Meta ad on Monday, doesn’t click, then searches for a product on Wednesday and converts with a paid product search ad. Meta records all your views. Google records last click conversions. A study of the lift in any channel alone may show a modest incremental contribution. The honest answer is that both ads did real work, just different work. Meta impressions created product consideration, while branded searches closed demand. Cutting anywhere breaks the trip.
This is exactly the conclusion many brands reach when they study the effects of scaling without proper context. They see Meta’s promotion research going down, conclude that the station is taking credit for conversions that would have happened anyway, and reallocate the budget. Six weeks later, product search volume is down, aggregate efficiency is down too, and the team is trying to figure out what happened.
One lift study at one station cannot tell you whether that station is worth the budget; it can only tell you what happened within the test, which is why allocation decisions need a metric that captures the whole business.
Marketing Efficiency Ratio (MER) Metric Discussion Not Available
The Marketing Efficiency Ratio, total revenue divided by total advertising spend, is the only commonly available metric that doesn’t care which channel gets the credit. It treats marketing as a single investment that produces a single return. That’s what marketing really is at the business level, and that’s the question CFOs and founders are asking when looking at performance.
MER alone is not enough. It will not tell you how to allocate within the budget, and it can be increased by the season or natural demand growth. But it does answer a question that should underpin all other valuation decisions: Does the combined marketing investment generate an acceptable return at the business level? Once that anchor is in place, the role of all the other layers becomes clear.
A Three Layer Stack That Really Works
A robust measurement stack has three layers, each answering a different question.
- MER Answers: Is the amount spent on marketing generating the returns this business needs? Does the investment work?
- Growth Answers: If I increase or decrease the amount spent on this channel, what happens to the MER?
- Attribution Answers: What touchpoints do customers engage with, and what does that tell me about the channel’s role? How does this affect the customer journey?
A mistake brands make is to use any one layer to answer questions that require more. Cutting Meta because product search closes the sale reads the adjective as a cause. Trusting Meta’s reported return on ad spend does the same thing in reverse. Treating the lift study as a decision on whether a station is eligible for disposal ignores whether that station may contribute to the MER through its effect on other stations.
How to Move Up Within This Stack
Upgrading testing was not as easy as it is now, and in some cases, the price tag was much higher than what they would want to invest. The good news is that the cost of conducting growth experiments has dropped significantly by 2025. Four test methods, ranked by accessibility:
Platform-Native Lift Studies
Meta Conversion Lift and Google Conversion Lift work within existing ad platforms at no additional cost. According to Google’s official Conversion Lift documentation, the platform is now reporting the results of a direct lift for courses with a budget of more than $5,000 USD and 1,000 conversions, based on the switch to a Bayesian statistical methodology that allows courses to work with lower budgets and fewer conversions than the old frequentist method required. Google Ads Highlights for 2025 confirm Conversion Lift is now operating at lower spend levels and conversion volume than in previous years.
Meta’s Brand Lift studies sit at the other end of the spending spectrum. According to Meta’s minimum requirements document, Brand Lift in the United States requires a minimum budget of $120,000 during the scholarship period. This is up from $30,000, which is a significant increase and puts Brand Lift out of reach for most companies. That said, Meta’s Conversion Lift courses have low limitations and are always a viable starting point. These two products measure different things and have very different costs, which should be understood before designing a test plan.
Native field testing has a clear limitation. They only measure elevation within the field running the test, so they cannot account for the results of different stations. Read the results as a single entry, not a decision.
Geo Holdout Test
If your marketing is spread across enough markets to continue to take hold, geo testing produces cleaner results than user optimization studies. Freeze money in the same markets while continuing in others, and measure the income gap. Test and control markets need to be matched on basic performance, seasonal patterns, and customer demographics, with several weeks of pre-test baseline data to ensure that the markets behave similarly under normal conditions.
Low Income Assessment
This is the most direct way to measure MER sensitivity. Cut channel budgets by 50 to 75% to get a defined window and measure total business impact, not channel-level metrics. If you cut the Meta in half and the total revenue is down 40%, that channel is contributing more than its lift research suggests. If you cut it in half and you have revenue, it’s possible that the channel was meeting the demand that other channels were creating. The low income test produces results that are less statistically robust than the well-planned geo holdout, but it is the only test that clearly measures the station’s contribution to MER rather than specified revenue.
Complete Models of Causal Inference
Manufacturing controls, analysis of variance, and experimental finite-media mixture modeling sit at the top of the methodological stack. Google’s open-source Meridian MMM, released in 2025, brought Bayesian causal inference modeling to marketers without requiring a proprietary marketing relationship, but the methodology still requires significant data science capabilities to implement effectively. Most brands don’t need to work at this stage to make defensible distribution decisions. The first three methods will answer important everyday budgeting questions.
Test Cadence That Creates a Real Signal
Effective cadence for a brand that spends $100,000 to $1 million monthly on paid channels:
- Weekly review of MER at aggregate level, broken down by new versus returning customer where possible.
- A quarterly increase test for the largest channel by spending, organized as a geo holdout where possible.
- Annual full-station catchment at each major station to update baseline estimates.
- Ongoing studies to promote the native place in new campaigns and important renewal of art.
- It’s a waste of time testing when the MER goes well without an obvious explanation.
Brands that build a quarterly testing rhythm develop a defensible view of channel sensitivity that no other platform dashboard can give them, and pairing that with a strong MER read sharpens the entire distribution conversation.
Learning to Increase Results Without Overcorrecting
The hardest part of growth testing is translating the results into context. Low lift research on Meta does not mean Meta should be cut. It means that the channel does not create an independent increase in volume during the evaluation window, the exception is that the channel submits MER for its effect on product search, direct traffic, or returning customers.
Read the lift lesson as one signal near the MER. If the elevator reaches the floor and the MER does not move when you reduce your spending, the channel can be changed. If the elevator reaches the ground but the MER goes down, the station performs an unmeasured test.
Stack Most Products Are Almost Built
Most growth stage brands have all three layers available to them and don’t use them as a stack. They look at the ROAS of the platform, periodically check the lift course, and treat the MER as a number that stays in the financial report instead of a valuation decision.
The growth debate has spent two years debating whether the attribute is broken. It is not a valid argument. Attribution describes the journey. The increase measures the sensitivity. MER is the metric by which a business is performing. Brands that build all three into a single decision-making process will allocate paid media budgets with greater confidence than those still debating which platform number to trust. If you don’t focus on MER and use expansion as a diagnostic to explain its movement, that’s a gap that needs to be closed first.
Additional resources:
Featured Image: Igor Link/Shutterstock
