Tangentix Work with us
Tangentix / Navigate
Work with us

Acquisition / Media Intelligence

Pay for the growth you actually caused.

You have more platform reports than you can read, and not one of them can tell you what your spend actually caused or where the next dollar belongs.

Which channels deserve the next dollar, and which don't?
Work with usReview the engagement

Allocation desk / planning range

Decision arena
Portfolio response / budget move
Evidence state
Modeled proposal / needs sign-off
TXMedia Optimizer
Illustrative framework / no observed outcomeWeek 12
Open evidence record Decision workspace / TX-WS-MED-01
Artifact
Decision workspace
Evidence status
Illustrative framework / no observed outcome
Record ID
TX-WS-MED-01
Record
Media Optimizer decision workspace / rolling 12 weeks / illustrative values
Decision question
Where should the next media dollar move, within what downside boundary?
Capital allocation / North America

Fund the next dollar against marginal return.

Rolling 12 weeks
Portfolio governed$4.8mannualized media
Proposed move$180kwithin portfolio
Modeled contribution+$220–390k80% planning range
Decision workspace · sanitized reconstruction · illustrative valuesSignal → funding decision → measurement rule

Engagement brief

One question, tightly drawn. One decision you can defend.

Built for
For the CMO, growth, media, and finance leaders who decide where the budget goes.
Start here
Media incrementality and budget-efficiency review
Decision record
A budget plan built on marginal return, with a built-in check on whether it worked
What we need from you
A weekly decision meeting plus access to your spend, outcome, and test data.

The analytical work

Evidence enters. A capital or operating decision leaves.

What we'll ask you for

  1. Channel, campaign, spend, and conversion history
  2. Your targets, margins, and constraints
  3. Any past tests, holdouts, and how you measure today

How we get to the answer

  1. Turn scattered platform reports into one trustworthy picture
  2. Estimate the baseline demand, then the growth you actually added
  3. Model where each channel stops earning, and how sure we are
  4. Turn it into budget options, each with its risk spelled out

Evidence sequence

Not dashboards for display. Evidence for the next move.

Every view below answers the three questions you'd ask anyway: what did we look at, what did we find, and what should you do about it?

TXMedia Optimizer
Illustrative framework / no observed outcomeWeek 12
Open evidence record Decision workspace / TX-WS-MED-01
Artifact
Decision workspace
Evidence status
Illustrative framework / no observed outcome
Record ID
TX-WS-MED-01
Record
Media Optimizer decision workspace / rolling 12 weeks / illustrative values
Decision question
Where should the next media dollar move, within what downside boundary?
Capital allocation / North America

Fund the next dollar against marginal return.

Rolling 12 weeks
Portfolio governed$4.8mannualized media
Proposed move$180kwithin portfolio
Modeled contribution+$220–390k80% planning range

Read every supporting value in the allocation record below.

ChannelCurrentMarginal indexSaturationProposal
Non-brand search28%1.42 58% +$105k
Paid social34%0.73 86% −$140k
Video prospecting16%1.18 49% +$75k
Affiliate22%0.91 76% −$40k
Capital state

A modeled move is not an authorization. An authorization is not a result.

We keep the forecast, the sign-off, and the after-the-fact review as separate records, so platform credit can't quietly turn into a budget decision.

  1. 01
    Modeled allocationProposal formed

    Baseline, 80% planning interval, marginal-response range, downside boundary, and exposure cap are ready for the decision forum.

    Proposed
  2. 02
    Authorization gateAwaiting accountable sign-off

    The move remains proposed until the decision owner and finance reviewer accept the guardrails.

    Not approved
  3. 03
    Realized reviewNo outcome observed

    After the review window, compare eligible outcomes with the frozen baseline before the next allocation moves.

    Not observed
Scenario desk

Compare the move before committing capital.

The ranges show model uncertainty. They're planning bounds, not promises.

Use horizontal scrolling to compare every allocation scenario.

ScenarioCapital movedContribution rangePrimary riskState
Hold current mix$0+$40k–$120kSaturation compoundsReference
Guardrailed reallocation$180k+$220k–$390kLearning preservedRecommended
Aggressive reallocation$340k+$250k–$470kThreshold at riskBounded
Baseline
Approved pre-move response model
Uncertainty basis
80% planning interval / marginal-response model
Review state
Pending realized outcome
Release condition
Lower bound remains positive after holdout or geographic adjustment
Decision workspace · sanitized reconstruction · illustrative valuesSignal → funding decision → measurement rule
Method record Illustrative framework / no observed outcome
01

Marginal return

See where average ROAS starts hiding waste.

The response curve shows you where your spend still earns, where it turns, and where it stops paying you back, before a single dollar moves.

Method figure Figure 01 / analytical interpretation This is a source deliverable for method inspection, not a product interface.
Open evidence record Research model / TX-METH-MED-01
Artifact
Research model
Evidence status
Illustrative framework / no observed outcome
Record ID
TX-METH-MED-01
Record
Marginal-return framework / illustrative values
Decision question
Where does your next media dollar still earn incremental return?
Source record / TX-METH-MED-01 Research method plate Curated source deliverable / method artifact, not a live product view
Tangentix marginal-return framework / illustrative values
What Tangentix found
Average ROAS quietly hid the point where every extra dollar started earning less.
Decision that follows
Fund the next dollar only where the return clears the agreed downside.
Decision state Illustrative framework / no observed outcome
02

The funding decision

Turn analysis into a weekly funding decision.

One page that spells out the move, the reasoning behind it, what we expect it to do, what could go wrong, and whose name is on it.

Open evidence record Decision workspace / TX-WS-MED-02
Artifact
Decision workspace
Evidence status
Illustrative framework / no observed outcome
Record ID
TX-WS-MED-02
Record
Weekly capital-decision framework / illustrative values
Decision question
What should move, who owns it, and what condition releases the capital?
Source record / TX-WS-MED-02 Decision workspace record Curated source deliverable / not a live product view
Tangentix decision-governance framework / illustrative values
What Tangentix found
Two channels looked equally efficient on the report. At the margin, they were nothing alike.
Decision that follows
Move budget within agreed limits, then measure what actually happened.

Investment committee record / MI-07

A recommendation is not an outcome.

The modeled upside, the sign-off, and the evidence of what actually happened stay in separate columns. That's how a forecast stops quietly turning into a result nobody checked.

Proposed · not approved
Open evidence record Executive brief / TX-BRIEF-MED-01
Artifact
Executive brief
Evidence status
Sanitized reconstruction
Record ID
TX-BRIEF-MED-01
Record
Sanitized investment committee record / proposal state / illustrative values
Decision question
What can you authorize today, and what evidence releases the next dollar?
Decision owner
CMO or growth executive · finance reviewer
Decision state

Three gates. Only one is open.

  1. 01Model record complete
    Proposal assembled

    The baseline, response model, bounded move, planning range, risks, and review design are frozen for the decision forum.

  2. 02Not approved
    Capital authorization

    The $180k reallocation is recommended with guardrails. It remains proposed, not approved, until the accountable buyer and finance reviewer sign.

  3. 03Not observed
    Realized-effect review

    After 14 days, the team reconciles eligible outcomes to the frozen baseline and records variance, interval, integrity, and the next funding action.

Authorization terms

Bound the capital before release.

Capital at risk
$180k reallocated within the existing portfolio
Modeled range
+$220–390k incremental contribution · 80% planning range
Uncertainty basis
Marginal-response model across the rolling 12-week planning window
Primary measure
Geo- or holdout-adjusted incremental contribution after variable product cost
Exclusions
Platform credit alone, ineligible geographies, known promotion shocks, and invalid assignment
Realized-outcome ledger

Blank by design until observation.

Forecast values are preserved beside realized values. Dashes mean no evidence, not zero impact.

MeasureFrozen forecastRealizedVarianceReview status
Working media moved $180k — — Awaiting authorization
Incremental contribution +$220–390k — — Pending measurement
Adjustment integrity Geo or holdout valid — — Pending QA
Next allocation Scale / hold / reverse — — No decision yet
Measurement boundary

This reconstructed memo demonstrates the governance deliverable. Values are illustrative and no client outcome is claimed.

Decision that follows

Authorize the bounded move, request revision, or decline. Scale is a later decision supported only by the realized record.

Method record Real methodology / anonymized evidence
03

Causal measurement

Design evidence that survives scrutiny.

Dose-response and comparison tests check whether the movement is really something you earned.

Method figure Figure 03 / analytical interpretation This is a source deliverable for method inspection, not a product interface.
Open evidence record Research model / TX-METH-INC-02
Artifact
Research model
Evidence status
Real methodology / anonymized evidence
Record ID
TX-METH-INC-02
Record
Sanitized methodology plate / client values withheld
Decision question
Does response strengthen with exposure in a way consistent with an earned effect?
Source record / TX-METH-INC-02 Research method plate Curated source deliverable / method artifact, not a live product view
Real anonymized methodology deliverable
What Tangentix found
There was a real effect. It was just a lot smaller than the credit the platform was claiming.
Decision that follows
Set the next investment against real evidence and confidence, not one ROAS number.

What your team receives

A budget plan your team can act on.

It lays out the baseline, where the return runs out, the recommendation, the owner, and how we'll check it worked, all before any budget moves.

  1. 01Incrementality test plan
  2. 02Marginal-return curves
  3. 03Budget-move options, with the risk on each

How we'll know it worked

  • Incremental revenue or contribution
  • Marginal return by channel
  • Overall efficiency and concentration risk
  • Forecast vs. what the move actually delivered

Start with media incrementality and budget-efficiency review

Put the next dollar under a better rule.

Bring the budget call you're arguing about. We'll find the baseline, the evidence, and the rule you need before more money moves.

Work with us

Capture record
Decision question
Presentation
What it doesn't prove
High-resolution evidence preview