TattvaDocs
Use casesPerformance marketing

Channel budget scenarios

"What if I move 10% from Meta to Google?" — projected revenue, CM3 and repeat-retention impact, with the assumptions stated.

Planned

This use case needs definitions that aren't in the Library yet (CAC). Until your data analyst adds them, Tattva can't answer it as shown here.

The questions: "If I increase Meta budget by 10%, what happens to revenue and CM3?"

"How do I become 30% less dependent on Meta?"

"If I scale acquisition by 30%, what does month-3 retention look like in six months?"

These are forecasting questions. They're hard to answer by hand because each one means modelling many scenarios. Ask them in Reasoning and Tattva builds the projection from your history, with its assumptions stated.

Who asks

  • The performance-marketing lead, in monthly planning
  • Leadership, during quarterly strategy
  • Founders, before a board meeting

Frequency

Monthly during planning, more often around quarterly strategy and budget revisions.

Data you need

  • Daily spend by channel × campaign type × product
  • Daily revenue (new and repeat) by channel × product
  • A customer table with first-purchase date (drives cohort retention)
  • Historical CM3 by acquisition channel and product
  • Results of incrementality tests where you have them (budget tests, geo holdouts)

Be realistic: without incrementality tests, a projection like this is directional, not precise. Treat it as decision support, not a promise.

How to ask it

A single change:

  • "What if I increase Meta spend by 10%? Forecast revenue, CAC and CM3 at 1, 3 and 6 months."

A reallocation:

  • "What if I move 50,000 a month from Meta acquisition to Google brand? Forecast the impact with a range."

A strategy:

  • "We're at X% Meta dependence. Show three ways to cut that by 30% over six months, with the revenue and CM3 trade-offs of each."

Retention:

  • "If I scale acquisition by 30% this quarter, what month-3 and month-6 repeat retention should I expect, based on current cohorts?"

What you'll get back

For each scenario, a good answer gives:

  1. Headline: the projected change in revenue, CAC and CM3, with a range.
  2. By horizon: estimates at 1, 3 and 6 months.
  3. By cohort: how new-customer retention carries through to lifetime value.
  4. Assumptions: what was modelled from your data and what was assumed.
  5. Where it's least sure: usually moves of more than 20% on a channel, or shifts into territory you've never tested.

If the answer doesn't state its assumptions, ask: "What did you assume, and which figures come from our data?" Open the Trail to see which figures were computed and which were projected.

The strategic question: "less Meta-dependent"

This one isn't a number; it's a strategy. Ask for:

  • three or four pathways (for example: grow Google brand, add a marketplace ad channel, build organic content, test influencer-led channels);
  • for each: the investment needed, the time to break even, how sensitive it is to Meta's costs changing, and the assumptions;
  • a side-by-side trade-off table you can read on one screen.

The logic tree

Channel reallocation question?
├── One channel's budget changes? → point forecast for that channel
├── Budget moves between channels? → forecast both, plus the retention effect
├── Reducing dependence on a channel? → several pathways with a trade-off table
└── Several moves at once? → run each leg, combine, show where they conflict

How to make it recurring

Most scenario questions are one-off. Two are worth repeating:

  • Monthly: "Last month we considered these scenarios. How did the actuals compare with the forecast?" This tells you how far to trust the forecasts.
  • Quarterly: a fresh "current strategy options" run during planning.

Ask them when the period closes and click Save as report on the answer. Reports aren't scheduled; they're the material for your planning meeting. Use the report's Export menu to print it to PDF or download it as Markdown.

Pitfalls

  • Treating a point forecast as exact. Read the range. "+5% to +15% revenue" is the answer, not "+10%".
  • Asking beyond your history. "What if I 5× Meta spend?" is outside anything in your data. The answer should say so; believe it.
  • Skipping incrementality. Forecasts drift without periodic tests. Plan one geo holdout or budget test a quarter.
  • Confusing a forecast with a promise. This is decision support.
Channel budget scenarios — Tattva Docs