---
title: "Marketing Science: Lift Testing and Measurement | MSG"
description: "Controlled market tests that show what your marketing caused: matched-market lift tests, channel interaction tests, budget optimization, segmentation, conjoint and MaxDiff."
---

[Skip to content](https://www.marketingsystemsguild.com/marketing-science#main)

[![Marketing Systems Guild. Marketing Mastered.](https://www.marketingsystemsguild.com/hs-fs/hubfs/raw_assets/public/MSG-Site-v2/msg-logo.png?width=1782&height=246&name=msg-logo.png)](https://www.marketingsystemsguild.com/?hsLang=en)

Menu

[Demand Generation](https://www.marketingsystemsguild.com/demand-generation?hsLang=en) [Marketing Science](https://www.marketingsystemsguild.com/marketing-science?hsLang=en) [Resources](https://www.marketingsystemsguild.com/resources?hsLang=en) [About](https://www.marketingsystemsguild.com/about?hsLang=en) [Contact](https://www.marketingsystemsguild.com/contact?hsLang=en) [Meet Vic](https://www.marketingsystemsguild.com/vic?hsLang=en)

[Home](https://www.marketingsystemsguild.com/?hsLang=en) / Marketing Science

Marketing Science services

# Know what your marketing actually caused.

Controlled tests in your own markets show which marketing made money. Then we show where the next dollar pays back most.

- Proof of what worked, not what the ad platforms claim
- A budget plan by channel and market, with the odds shown
- Answers your CFO can sign off on

[Book a call](https://www.marketingsystemsguild.com/contact?hsLang=en) [See what you get](https://www.marketingsystemsguild.com/marketing-science#what)

How a lift test reads

Sales over the same weeks

Test markets  
change appliedBaselineLift

Control markets  
no changeBaseline

**The gap is the lift.** Same kinds of markets, same weeks. The only difference is the change you made.

Illustration, not client data.

What you get

## What can Marketing Science tell you?

Five questions every growing company asks. We answer each one with numbers you can act on.

1

### What actually worked?

We give a change to some markets and hold it back from others. The difference shows what your marketing added. It works for pricing, promotions and store hours too.

**You get:** Proof of what each change caused.

2

### What is each channel really returning?

We show how each channel's results build over time, where extra spend stops paying off, and how channels help each other. It works even when tracking is spotty.

**You get:** A return curve for every channel.

3

### Where should the next dollar go?

We run thousands of what-if scenarios on your budget to show where more money earns the most, and how sure we are.

**You get:** A budget plan by channel and market, with the odds.

4

### Which customers are worth the most?

We estimate what each customer is worth over time, who is likely to leave, and how they react to price. Then we group customers and markets by how they behave.

**You get:** The customers and markets to put money behind.

5

### How do buyers choose?

We test how buyers weigh features, price and brand before you build, price or promote.

**You get:** What buyers value most, ranked.

### Not sure where to start?

Bring one question you can't answer today. We'll tell you which test answers it and what it takes.

[Book a call](https://www.marketingsystemsguild.com/contact?hsLang=en)

How to start

## How do you start?

Three steps. Each one pays for itself before you commit to the next.

1. 1
   
   ### Test
   
   We take one decision you're about to spend money on and run it against matched control markets.
   
   **You get:** what it actually returned, with a confidence range.
2. 2
   
   ### Prove
   
   We build the testing system inside your business and map how each channel responds to spend.
   
   **You get:** a budget plan your CFO can check line by line.
3. 3
   
   ### Scale
   
   Results go stale when markets change, so we retest every quarter. We train your team and coach the marketing leader already in the seat.
   
   **You get:** a team that runs the method without us.

[Start with a test](https://www.marketingsystemsguild.com/contact?hsLang=en)

[Watch the video](https://www.marketingsystemsguild.com/hubfs/msg-site-v2/budget-cfo-yes-stopmotion.mp4?hsLang=en)

A real-world example · 3 min 42 sec

## How do you get a CFO to say yes to a marketing budget?

Bring the odds, not the average. This video works through the budget question every marketing leader faces, the Marketing Science way.

- Why a plan built on last year's average comes up short
- How small markets borrow strength from the rest, so one lucky month can't fool you
- How a $500K plan plays out 10,000 possible quarters, so you see the odds of beating break-even and the worst case before money goes out

The numbers in the video are an example, not client data.

[Stress-test your budget with us](https://www.marketingsystemsguild.com/contact?hsLang=en)

The method

## How does a test run?

1. 1
   
   ### Check feasibility
   
   We confirm you have enough markets and enough data to read a result before anything launches.
2. 2
   
   ### Group your markets
   
   Cluster analysis groups markets that behave alike.
3. 3
   
   ### Match test and control
   
   Every test market gets a matched control market from the same group.
4. 4
   
   ### Design the test
   
   One change at a time, or several at once with a factorial design. Tests can be budget-neutral: spend shifts between markets, it doesn't grow, and no market goes dark.
5. 5
   
   ### Run and read
   
   Weekly results while it runs, and a final read with its confidence range. You see the test design before launch and the result after.
6. 6
   
   ### Act and retest
   
   Budget moves to what the test proved. We retest every quarter, because a result holds only as long as market conditions hold.

Who it's for

## Is this for you?

If you sell in more than a handful of markets or locations, you can test this way. That includes:

- Multi-location operators and franchisees
- Regional retail chains
- Lead-generation businesses that close sales offline in a CRM
- Marketing leaders who need answers the CFO accepts

The first step checks whether you have enough markets to read a result, before you spend money on a test.

Your data

### What happens to your data?

Your raw data runs only through code. AI tools never see it. AI reads only the summarized results. If a file arrives with personal information in it, we send it back and tell you. We use it only with your permission, after stripping that information out. Client files are deleted within 30 days after the engagement ends.

The definition

## What is marketing science?

Marketing science uses the methods of science to make marketing decisions. Instead of gut feel or vanity numbers, it treats marketing as cause and effect: change one thing, measure what happens, and put the money where it pays back.

### How is it different from marketing analytics?

| Dimension | Marketing analytics | Marketing science |
| --- | --- | --- |
| The question | What happened? | Why did it happen, and what happens if we change it? |
| The tools | Dashboards and reports | Statistical models and controlled tests |
| The focus | Tracking the numbers | Finding what caused the numbers |
| The output | A report on last quarter | A plan for where to spend next |

**Analytics tells you what happened. Marketing science tells you what to do next.**

Common questions

## What else should you know?

### What is marketing science?

Marketing science uses the methods of science to make marketing decisions. It treats marketing as cause and effect: change one thing, measure what happens, and put the money where it pays back.

### How is marketing science different from marketing analytics?

Analytics reports what happened, using dashboards and reports. Marketing science finds out why it happened and what will happen if you change something, using statistical models and controlled tests. Analytics gives you a report on last quarter. Marketing science gives you a plan for where to spend next.

### What is a matched-market lift test?

A controlled experiment run in real markets. Test markets get a change, such as more spend in one channel. Matched control markets, chosen because they behave alike, don't. The difference in sales between the two groups is the lift the change caused.

### How do you tell marketing-generated demand from demand that would have arrived anyway?

We hold the marketing back from some markets and compare. First we sort your markets into groups that behave alike: size, sales history, season. From each group we pick test markets and control markets at random. Test markets get the marketing. Control markets don't. Control shows the demand that would have arrived anyway. The gap between test and control is the demand your marketing created.

### How do you measure incrementality when sales cycles are long and channels interact?

Each group of alike markets runs its own test against its own control, so a quirk in one group can't fake a result. Then we combine every group's result into one read on total impact, with its confidence range. For long sales cycles, the test runs long enough to cover the cycle, and we track qualified meetings and pipeline against control until closed revenue lands. For channels that interact, we change them together in some markets and apart in others. The test shows what each channel adds alone and what they add together.

### How is this different from marketing mix modeling?

Marketing mix modeling estimates what worked from past spend and sales. A lift test measures it directly with a controlled experiment. They work together: lift test results calibrate a mix model. We build both.

### Why retest every quarter?

A test result holds only as long as market conditions hold. Competitors, prices and seasons change. Quarterly retests and model recalibration keep your budget decisions current.

### Will any of my markets go dark during a test?

No. Tests can be budget-neutral. Spend shifts between markets, total spend doesn't grow, and no market goes dark.

### What does it cost?

Pricing is set to the test design: how many markets, how many changes and how long it runs. You can start with a single test on one question, so the first step stays small.

### Do I need a big budget to run a test?

No. Tests can be budget-neutral. Spend shifts between markets, it doesn't grow, so you test with the money you already spend.

### Who runs the campaigns during a test?

Your team runs the campaigns in your own ad platforms and keeps full say on creative. We design the test, set the budget levels that make it readable, and run the analysis.

Start here

## Bring one question you can't answer today.

Which channel is really driving sales? Where should the next dollar go? We'll tell you how a test would answer it.

[Book a call](https://www.marketingsystemsguild.com/contact?hsLang=en)

![Marketing Systems Guild. Marketing Mastered.](https://www.marketingsystemsguild.com/hs-fs/hubfs/raw_assets/public/MSG-Site-v2/msg-logo-dark.png?width=3602&height=538&name=msg-logo-dark.png)

Build it. Prove it. Scale it.

Marketing Systems Guild LLC  
New Jersey, USA  
[hello@marketingsystemsguild.com](mailto:hello@marketingsystemsguild.com)  
[215.858.0727](tel:+12158580727)

## Services

[Demand Generation](https://www.marketingsystemsguild.com/demand-generation?hsLang=en) [Marketing Science](https://www.marketingsystemsguild.com/marketing-science?hsLang=en)

## Start

[Meet Vic](https://www.marketingsystemsguild.com/vic?hsLang=en) [Contact us](https://www.marketingsystemsguild.com/contact?hsLang=en) [The Cold Take newsletter](https://www.marketingsystemsguild.com/the-cold-take?hsLang=en) [Subscribe on Substack](https://billbecht.io/subscribe)

## Learn

[Resources](https://www.marketingsystemsguild.com/resources?hsLang=en) [Questions and answers](https://www.marketingsystemsguild.com/resources#questions) [The Demand Engine Playbook](https://www.marketingsystemsguild.com/demand-engine-playbook?hsLang=en) [Capture vs creation](https://www.marketingsystemsguild.com/capture-vs-creation?hsLang=en)

## Company

[About](https://www.marketingsystemsguild.com/about?hsLang=en) [Privacy policy](https://www.marketingsystemsguild.com/privacy-policy?hsLang=en)

© 2026 Marketing Systems Guild LLC

![](https://www.marketingsystemsguild.com/hs-fs/hubfs/raw_assets/public/MSG-Site-v2/vic-pebble-fab.png?width=80&height=80&name=vic-pebble-fab.png)![](https://www.marketingsystemsguild.com/hs-fs/hubfs/raw_assets/public/MSG-Site-v2/vic-pebble-fab-wave.png?width=80&height=80&name=vic-pebble-fab-wave.png) Talk to Vic

Vic · Free · A few minutes

## What should you fix first?

Tell Vic how your marketing runs today, in your own words. No prep needed. You get your top finding and two fixes you can make this week.

Work email

Start with Vic

Our team follows up at this address. [Privacy policy](https://www.marketingsystemsguild.com/privacy-policy?hsLang=en). [What is Vic?](https://www.marketingsystemsguild.com/vic?hsLang=en)

```json
{
  "@context" : "https://schema.org",
  "@id" : "https://www.marketingsystemsguild.com/#org",
  "@type" : "ProfessionalService",
  "address" : {
    "@type" : "PostalAddress",
    "addressCountry" : "US",
    "addressRegion" : "NJ"
  },
  "alternateName" : "MSG",
  "areaServed" : {
    "@type" : "Country",
    "name" : "United States"
  },
  "description" : "Marketing Systems Guild is a boutique operator that builds and runs demand infrastructure inside its clients' businesses, then hands it to their teams. Two services: Demand Generation, which builds the system that finds buyers and brings them in, and Marketing Science, which runs controlled market tests that show which marketing makes money.",
  "email" : "hello@marketingsystemsguild.com",
  "founder" : [ {
    "@id" : "https://www.marketingsystemsguild.com/about#bill-becht",
    "@type" : "Person",
    "name" : "Bill Becht"
  }, {
    "@id" : "https://www.marketingsystemsguild.com/about#heidi-becht",
    "@type" : "Person",
    "name" : "Heidi Becht"
  } ],
  "knowsAbout" : [ "Demand generation", "Marketing measurement", "Matched-market testing", "Incrementality testing", "Marketing mix modeling", "Customer segmentation", "Conjoint analysis", "MaxDiff", "HubSpot", "Lead scoring" ],
  "legalName" : "Marketing Systems Guild LLC",
  "logo" : "https://www.marketingsystemsguild.com/hubfs/raw_assets/public/MSG-Site/msg-logo.png",
  "makesOffer" : [ {
    "@type" : "Offer",
    "itemOffered" : {
      "@type" : "Service",
      "name" : "Pipeline Blueprint"
    },
    "name" : "Pipeline Blueprint",
    "price" : "7500",
    "priceCurrency" : "USD",
    "url" : "https://www.marketingsystemsguild.com/demand-generation#pricing"
  } ],
  "name" : "Marketing Systems Guild",
  "sameAs" : [ "https://www.linkedin.com/in/billbecht" ],
  "slogan" : "Build it. Prove it. Scale it.",
  "telephone" : "+1-215-858-0727",
  "url" : "https://www.marketingsystemsguild.com/"
}
```

```json
{
  "@context" : "https://schema.org",
  "@type" : "BreadcrumbList",
  "itemListElement" : [ {
    "@type" : "ListItem",
    "item" : "https://www.marketingsystemsguild.com/",
    "name" : "Home",
    "position" : 1
  }, {
    "@type" : "ListItem",
    "item" : "https://www.marketingsystemsguild.com/marketing-science",
    "name" : "Marketing Science",
    "position" : 2
  } ]
}
```

```json
{
  "@context" : "https://schema.org",
  "@type" : "FAQPage",
  "mainEntity" : [ {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "Marketing science uses the methods of science to make marketing decisions. It treats marketing as cause and effect: change one thing, measure what happens, and put the money where it pays back."
    },
    "name" : "What is marketing science?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "Analytics reports what happened, using dashboards and reports. Marketing science finds out why it happened and what will happen if you change something, using statistical models and controlled tests. Analytics gives you a report on last quarter. Marketing science gives you a plan for where to spend next."
    },
    "name" : "How is marketing science different from marketing analytics?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "A controlled experiment run in real markets. Test markets get a change, such as more spend in one channel. Matched control markets, chosen because they behave alike, don't. The difference in sales between the two groups is the lift the change caused."
    },
    "name" : "What is a matched-market lift test?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "We hold the marketing back from some markets and compare. First we sort your markets into groups that behave alike: size, sales history, season. From each group we pick test markets and control markets at random. Test markets get the marketing. Control markets don't. Control shows the demand that would have arrived anyway. The gap between test and control is the demand your marketing created."
    },
    "name" : "How do you tell marketing-generated demand from demand that would have arrived anyway?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "Each group of alike markets runs its own test against its own control, so a quirk in one group can't fake a result. Then we combine every group's result into one read on total impact, with its confidence range. For long sales cycles, the test runs long enough to cover the cycle, and we track qualified meetings and pipeline against control until closed revenue lands. For channels that interact, we change them together in some markets and apart in others. The test shows what each channel adds alone and what they add together."
    },
    "name" : "How do you measure incrementality when sales cycles are long and channels interact?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "Marketing mix modeling estimates what worked from past spend and sales. A lift test measures it directly with a controlled experiment. They work together: lift test results calibrate a mix model. We build both."
    },
    "name" : "How is this different from marketing mix modeling?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "A test result holds only as long as market conditions hold. Competitors, prices and seasons change. Quarterly retests and model recalibration keep your budget decisions current."
    },
    "name" : "Why retest every quarter?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "No. Tests can be budget-neutral. Spend shifts between markets, total spend doesn't grow, and no market goes dark."
    },
    "name" : "Will any of my markets go dark during a test?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "Pricing is set to the test design: how many markets, how many changes and how long it runs. You can start with a single test on one question, so the first step stays small."
    },
    "name" : "What does it cost?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "No. Tests can be budget-neutral. Spend shifts between markets, it doesn't grow, so you test with the money you already spend."
    },
    "name" : "Do I need a big budget to run a test?"
  }, {
    "@type" : "Question",
    "acceptedAnswer" : {
      "@type" : "Answer",
      "text" : "Your team runs the campaigns in your own ad platforms and keeps full say on creative. We design the test, set the budget levels that make it readable, and run the analysis."
    },
    "name" : "Who runs the campaigns during a test?"
  } ]
}
```

```json
{
  "@context" : "https://schema.org",
  "@type" : "Service",
  "areaServed" : {
    "@type" : "Country",
    "name" : "United States"
  },
  "description" : "MSG runs matched-market lift tests, channel interaction tests, response curves for budget optimization, customer and market segmentation, and conjoint and MaxDiff research.",
  "name" : "Marketing Science",
  "provider" : {
    "@id" : "https://www.marketingsystemsguild.com/#org"
  },
  "serviceType" : "Marketing measurement and incrementality testing",
  "url" : "https://www.marketingsystemsguild.com/marketing-science"
}
```

```json
{
  "@context" : "https://schema.org",
  "@type" : "VideoObject",
  "contentUrl" : "https://245907537.fs1.hubspotusercontent-na2.net/hubfs/245907537/msg-site-v2/budget-cfo-yes-stopmotion.mp4",
  "description" : "Bring the odds, not the average. A short stop-motion explainer from Marketing Systems Guild: hierarchical Bayesian regression and a 10,000-quarter simulation turn a single budget forecast into odds a CFO can underwrite. Example numbers, not client data.",
  "duration" : "PT3M42S",
  "embedUrl" : "https://www.marketingsystemsguild.com/marketing-science#example",
  "name" : "How do you get a CFO to say yes to a marketing budget?",
  "publisher" : {
    "@id" : "https://www.marketingsystemsguild.com/#org"
  },
  "thumbnailUrl" : "https://245907537.fs1.hubspotusercontent-na2.net/hubfs/245907537/msg-site-v2/budget-video-poster.jpeg",
  "uploadDate" : "2026-10-19"
}
```