Cassandra vs Haus: an incrementality platform, or an AI marketing brain that does the work
Haus calls itself an AI-powered incrementality platform: experiments, a Causal MMM it says is built on them, and Causal Attribution. Cassandra, an AI marketing brain for growth teams, measures with a Bayesian marketing mix model checked by live geo experiments, then does the work.
Both feed experiment results into a marketing mix model. Haus says its model treats experiments as ground truth; Cassandra starts from a model of the whole mix, checked by experiments.
What Cassandra is
- An AI marketing brain for growth teams that learns what drives growth and does the work.
- Measures with a Bayesian marketing mix model checked by live geo experiments.
- Does the work within the rules the team sets: plans, budget moves, ads, creative and reports.
- Approvals: the team decides what waits for approval, and every move is checked against its forecast.
How this comparison was made
Everything about Haus comes from its own public pages, read on 12 October 2026. We have not tested their product.
If anything is out of date, write to hello@cassandra.app and we will correct it.
When Haus is the better choice
- Experiments first: Haus leads with experiments, including GeoLift, Time Tests and Fixed Geo Tests.
- A strategist: Core and above include a dedicated Measurement Strategist; Enterprise adds a data scientist.
- Self-serve: Haus describes its Basics plan as self-serve for DTC brands selling via Shopify and Amazon.
- Daily adjusted reporting: Causal Attribution adjusts platform reporting based on experiment results and refreshes daily.
What each one says it is
- Haus describes itself as "The AI-powered incrementality platform leading enterprises use to optimize tens of billions in annual marketing spend".
- Its products: Incrementality Experiments, a weekly Causal MMM, Causal Attribution, and Architect, which it calls a causal marketing agent.
- Cassandra describes itself as an AI marketing brain for growth teams.
Where the claim sets differ
- Where each starts: Haus says tests drive its model's structure; Cassandra starts from a Bayesian model of the mix.
- Attribution: Haus offers Causal Attribution; Cassandra checks attribution against measured lift.
- After the measurement: Haus offers Architect; Cassandra plans, moves budgets, manages ads, makes creative and writes reports.
- Ask each exactly how a new result changes the model.
What a test tells you, and what a model adds
- A geo test measures one channel, in one window, at one spend level.
- A model carries it across the mix: fixing one channel's effect moves the channels it overlaps with.
- Example: a test finds paid social returns less than assumed; the estimate for display, whose spend moved with it, shifts too.
What Haus publishes about pricing
- Four plans: Basics for Shopify, Core, Plus and Enterprise; Plus and Enterprise list Causal MMM.
- Quoted on request: each plan has a "Get a demo" button (as listed on 12 October 2026).
- Cassandra: start for free with your website and data. Paid plans scale with yearly ad spend.
How each fits a testing programme
- Testing-first: tests run continuously and a model built on them fills the gaps. Haus's published emphasis fits this.
- Model-first: a model covers the mix; tests go where it is least certain. Cassandra fits this.
Limits worth knowing
- Spend history: a channel needs enough spend, and variation in it, for the model to read.
- Regions: geo experiments need regional structure to build a comparison group.
- Time: tests need a few weeks, free of other changes in the test regions.
Cassandra is an AI marketing brain for growth teams. It measures with a Bayesian marketing mix model checked by live geo experiments, and does the work. Start for free with your website and data, or book a demo.
What moving between platforms involves
- The experiment record moves with you: each test stays valid evidence in any model.
- So does the cleaned history of spend and outcomes; what changes is the programme's rhythm.
- Run both on the same period and compare how each uses the same recent test.
How to choose
- Haus: experiments as the foundation, with a strategist and a model built on them.
- Cassandra: a model of the whole mix, checked by geo experiments, and the work done for the team.
- Daily experiment-adjusted platform reporting: Haus offers Causal Attribution.
Questions
What is the difference between Cassandra and Haus for incrementality and MMM?
Both run geo experiments and feed the results into a marketing mix model. Haus says its Causal MMM is built on an experimental foundation. Cassandra, an AI marketing brain, checks a Bayesian model with experiments and does the work: plans, budgets, ads, creative and reports.
How is Haus priced?
Haus's pricing page lists four plans, Basics for Shopify, Core, Plus and Enterprise, each with a "Get a demo" button, so pricing is quoted on request (as listed on 12 October 2026).
How does Haus use experiments in its MMM?
Haus says it treats experiments as ground truth in its Causal MMM, starting with experiments and letting observational data fill in the gaps. It says the model is updated weekly.
Does Haus publish a methodology?
Yes. Haus's Science page says its products are rooted in explainable econometrics and causal inference, and describes synthetic controls and on-demand power analyses.
What are alternatives to Haus?
Vendors that publish geo experiments include Measured, Sellforte and SegmentStream. Cassandra, an AI marketing brain, is another. Our list of incrementality tools sets out who each suits.
Is Cassandra a good Haus alternative for ecommerce brands?
It can be, depending on the team. Both measure with geo experiments and a marketing mix model; Cassandra also moves budgets, adjusts ads and makes new ones. Ecommerce brands such as Gina Tricot and Velasca use it. Haus lists a Basics plan for Shopify.