T2K in practice / Real estate marketing
A postcard, a search, a phone call. One property acquisition. Which marketing deserves credit—and what supports the next budget?
Follow the evidenceInteractive showcase · Fictional operator · Synthetic data
The case study
A fictional regional operator buys houses from homeowners, renovates them, and resells them. T2K connects the evidence needed to review a practical question: should the next marketing budget move money from direct mail to paid search?
A transparent demonstration. MTA calculations use a synthetic event ledger. Google Meridian is fitted to generated weekly data; a model diagnostic requires review. These are synthetic examples, not real performance results. Budget mixes are authored examples, not optimized recommendations; the final view checks their bounds and operating capacity without changing advertising spend.
Read the guide01 / Explain this acquisition
Multi-touch attribution (MTA) assigns credit under a declared policy. Change the evidence and see what still supports the answer.
Loading the public measurement bundle…
Each event keeps its date, source, and evidentiary meaning.
The selected snapshot determines which records were available.
Eligible interactions are evaluated against the selected policy and reporting cutoff.
Loading credit allocations…
Credit explains an allocation. It does not establish causal lift.
Purchase capital, projected profit, and realized contribution remain separate.
Source identifiers, exclusions, the cutoff, and the policy accompany each result. The application calculates attribution fractions. The separate Core receipt records the eligible evidence query.
Loading…
Loading…Open the complete synthetic measurement bundle (JSON)
02 / Assess channel investment
Marketing mix modeling (MMM) estimates outcomes that would change with marketing investment. That is a different question from assigning credit to this property.
The first useful outcome is newly qualified seller-property opportunities. Completed acquisitions require their own data sufficiency assessment.
| Channel | Spend | Incremental outcomes | Cost per incremental outcome |
|---|
A common outcome, period, geography, and channel definition are necessary before results can be compared.
This synthetic fit requires model review. It does not validate real-world marketing effectiveness or a spending recommendation.
03 / Review the next decision
Bring attribution, model uncertainty, and operating limits into one review. Acquisition capital and assessment capacity matter alongside channel performance.
Try a bounded budget proposal
Compare an authored monthly spending mix with channel bounds, acquisition capital, assessments, and renovation capacity. These are planning assumptions, not model forecasts.
No campaign is updated. The result records what still needs review.
The selected spending mix, measurement evidence, and operating constraints will appear here.
The selected MMM outcome also determines what model evidence is available.
These review inputs preserve the selected journey, model mode and readiness, operating constraints, and lack of spending authority. This is a showcase review, not a production authorization record.
Loading…
Methods and limits
The data, calculations, and limitations are available for inspection. This is a technical demonstration of T2K, not a customer performance claim.
MTA explains an allocation of credit; MMM estimates incremental effects conditional on a model. T2K Core records the evidence eligibility query, while application code calculates credit, economics, and capacity. The 2026 property journey and the aggregate MMM dataset—156 weeks from 2023 to 2025 across four fictional markets—are separate examples, not a joined customer funnel.
Meridian’s sampling checks pass, but its 90% baseline interval crosses below zero and 5.3% of retained draws have a negative aggregate baseline. The model status is review_required. No holdout evaluation or experiment calibration was performed. The sparse acquisition series is insufficient_data; no acquisition or profit model was fitted.
No. The budget mixes are human-authored scenarios, with no future forecast or budget optimization. Passing channel and capacity checks does not resolve model uncertainty or authorize spending. Purchase capital also excludes renovation, holding, and marketing funding, which require separate review.
Inspect the full methodology and reproduction guide and model artifact. The implementation uses Google’s Meridian data requirements and model execution documentation.
What T2K brings together
The same structure can connect a seller journey, a measurement model, and a reviewed business decision. Every result keeps its meaning and its limitations.
Build with T2K