Skip to main content
Evidence. Review. Action.
Case study

Cash Homebuyer Marketing Attribution: MMM + MTA

Connect seller journeys, channel evidence, and operating capacity before deciding where the next marketing dollar should go.

By sig.ai

THE DECISION

Should a regional homebuyer move budget from direct mail to paid search?

Try the evidence rules, inspect the model, and test the constraints in T2K.

Explore the interactive T2K demo

Fictional people, properties, and results. All amounts and datasets are synthetic.

01 / BUSINESS CONTEXT

Measure through to the property.

A regional operator that buys, renovates, and resells houses needs qualified sellers, viable properties, and enough capital to close. Cheap inquiries alone cannot answer whether marketing supports the business. Direct mail, search ads, social campaigns, and referrals can all appear along the same seller journey.

The measurement unit is a seller-property opportunity. One property may have two owners, repeated calls, and a contract that later falls through. The operator needs consistent identities, source records, milestone dates, and complete costs before comparing channels or reporting realized contribution.

  1. 1Inquiry
  2. 2Qualified opportunity
  3. 3Assessment
  4. 4Offer
  5. 5Contract
  6. 6Acquisition
  7. 7Resale

02 / TWO MEASUREMENT QUESTIONS

Where MTA and MMM fit

Multi-touch attribution

Which eligible sources receive credit?

MTA follows recorded interactions. First touch, last touch, and equal credit across distinct eligible channels produce different allocations under declared rules. The outcome here is a newly qualified seller-property opportunity. Those allocations express a credit policy; they do not establish causal lift.

Marketing mix modeling

What supports channel investment?

An actual Google Meridian fit uses 156 generated weekly observations per market across four fictional markets. It estimates qualified opportunities associated with direct mail, paid search, and paid social, conditional on the model. Its 90% posterior intervals expose uncertainty; brand and nonbrand search are combined.

The aggregate model and individual seller journey are separate synthetic examples. Their totals cannot be joined into a lead-to-close conversion rate. Organic and referral effects are not separately estimated.

03 / EVIDENCE BEFORE ALLOCATION

What changes when the record changes?

In the demo, a postcard dispatch proves mailing, not attention. A reviewed seller statement can add reported influence under a separate policy. Duplicate call deliveries, coowner callbacks, and follow-up reminders do not create extra property acquisitions. Missing call-source mapping withholds final allocation while preserving independently supported clicks.

Changing the reporting cutoff shows only evidence available at that time. A canceled contract remains a historical contract with zero acquisitions. T2K Core executes the evidence eligibility query; application code calculates credit, economics, and capacity. The downloadable receipts let reviewers inspect that boundary.

An illustrative property ledger

A $295,000 resale less a $210,000 purchase, $40,000 renovation, and $18,000 holding, financing, and transaction costs leaves $27,000 contribution before marketing. A declared $3,000 marketing allocation reduces it to $24,000. Before resale or with missing costs, realized contribution is unknown. These amounts are neither benchmarks nor a forecast.

04 / A REVIEWABLE DECISION

A feasible budget still needs judgment.

Keep the model limitation visible

Sampling diagnostics pass, but the fitted baseline interval crosses below zero. The model therefore requires review. The sparse acquisition series has no fitted model, and there is no profit model, holdout evaluation, experiment calibration, or validation with a real operator.

Check the business constraints

Three human-authored $90,000 monthly budget mixes test channel bounds, purchase capital, assessment slots, and renovation capacity. One moves $9,000 from mail to search. These are planning examples without optimization or a forecast. Renovation, holding, and marketing funding need separate review. No advertising account is connected or spending authorized.

05 / OPEN THE EVIDENCE

Inspect the methods

Start with a seller journey, change an assumption, then inspect what supports the result. The public artifacts document the data, rules, model fit, and limits of this example.

Explore the interactive T2K demo