How we work

Accountable judgment. Close to the evidence.

The people who frame your question, test the evidence, and defend the recommendation are the same people you'll be sitting across from.

No account managers. Nothing gets lost in a handoff between your problem and the analysis, because there isn't one.

Decision rights in the room

Every judgment has an accountable owner.

Three non-delegable decision rights, assigned by name for each engagement. This is not a headcount and not a staffing pyramid: each right means one named person owns a specific call, sits in a specific forum, and puts their name on a specific deliverable.

Engagement role

Tangentix engagement lead

Accountable judgmentCommercial framing and the final recommendation

Non-delegable ruleCannot hand final sign-off on the recommendation to anyone else.

Present in: Opening brief · decision forums · executive review
Signs off: Decision brief · investment recommendation · executive narrative

Engagement role

Quantitative lead

Accountable judgmentCounterfactual, model design, assumptions, and uncertainty

Non-delegable ruleMust tell you plainly what the evidence will not support.

Present in: Evidence audit · model review · measurement gate
Signs off: Baseline · sensitivity analysis · causal test plan

Engagement role

Product & operations lead

Accountable judgmentTranslation from recommendation into an operating system

Non-delegable ruleNothing ships without someone's name on it and a way to measure it.

Present in: Working sessions · owner reviews · operating cadence
Signs off: Prioritized roadmap · decision rights · review protocol

One person can hold more than one right on an engagement, and no right is ever split between two. You receive the proposed named engagement team, and which right each person holds, before you agree scope.

How we make a recommendation

A recommendation does not move because it is persuasive.

Plenty of confident-sounding advice has cost people their budgets. Ours moves only when the record behind it is complete enough for your team to challenge it, run it, and check afterwards whether it worked.

01Frame review

Name the decision before selecting the method.

Before anyone touches the data, we write down what you're deciding, what we're comparing it against, who owns the call, and what it costs you to get it wrong.

Review question
What decision is this evidence allowed to change?

Record
Decision frame and boundary
02Evidence review

Make the source record inspectable.

You get to see what went in, what we left out, where the data thins out, and which conflicts we couldn't resolve. No result gets trusted on faith.

Review question
What can the evidence support, and what is still unknown?

Record
Evidence inventory and limitation note
03Challenge review

Test the explanation, not just the movement.

We argue against our own finding first: what else could explain this, how bad is the downside, and is there a fairer comparison we should be using?

Review question
What else could plausibly explain the observed result?

Record
Challenge note and decision rule
04Release review

Assign the action and the rule that governs it.

Nothing ships without a name against it, a way to measure it, a date to review it, and an agreed rule for when to pause, scale, or stop.

Review question
What would have to be true for the decision to survive review?

Record
Action release and realized-review plan

How we engage, and what it costsWhat the first 90 days look like, and how we price the work.

The first 90 days

Three handoffs turn the question into an answer you can act on.

You'll never be a month in wondering what you're paying for. Every 30 days ends with something real in your hands, a name against it, and a gate to clear. We don't move on until the decision is genuinely ready.

01 Days 1–30

Decision question

Establish the truth
What will the evidence actually support today, and where is the baseline still shaky?

Primary handoffTX / BASELINE-01

Decision brief
Baseline and evidence boundary

Decision question · evidence inventory · baseline definition · known limits

StatusReady for approval

What we need from you
Two working sessions · governed data access · decision owner named

Exit gate Baseline and evidence limits approved
02 Days 31–60

Decision question

Set the next move
Which opportunity survives competing explanations and deserves action first?

Primary handoffTX / DECISION-02

Recommendation brief
Opportunity and scenario record

Ranked opportunity · competing explanation · downside boundary · recommended move

StatusDecision-ready

What we need from you
Weekly decision forum · assumption review · operating constraints

Exit gate Recommendation and measurement rule signed off
03 Days 61–90

Decision question

Operate the answer
How does this go live, earn the right to scale, and get checked against what really happened?

Primary handoffTX / REALIZED-03

Outcome review
Realized decision record

Operating plan · launch gate · observed variance · next funding rule

StatusOperating

What we need from you
Execution owner · test integrity · outcome instrumentation

Exit gate Next allocation tied to realized evidence

Measurement and pricing

We price the work around the decision.

What you pay depends on how hard the decision is, how good your evidence already is, and how often we're in the room. It has nothing to do with how much you spend on media, how many dashboard seats you buy, or a promise of certainty nobody can honestly make.

01Fixed scope

Recommended first engagement

Bounded diagnostic

One big call is coming, and you want solid ground under it before you commit.

Cadence: Typically 4–8 weeks
Includes: Decision brief · evidence audit · analysis · recommendation · measurement plan
Buyer commitment: One accountable decision owner, governed evidence access, and two working sessions.
Measurement: Decision readiness and an agreed next proof move
Release condition: The decision record names the next proof move, owner, and measurement condition.

02Defined cadence

For recurring decisions

Operating partnership

The same decision keeps coming back every quarter and deserves a proper rhythm.

Cadence: 90-day operating cycles
Includes: Decision forum · evolving models · scenarios · operating backlog · realized review
Buyer commitment: A recurring decision forum, an execution owner, and access to the operating evidence.
Measurement: Forecast versus realized outcome and the quality of the next allocation
Release condition: The owners accept the operating backlog, the review cadence, and how we'll check what it earned.

03Base + qualified variable

Only after baseline approval

Outcome-linked structure

You want our fee tied to the outcome, and you can agree the baseline, the measure, the data, the timing, and who does what up front.

Cadence: Only where evidence permits
Includes: Pre-agreed counterfactual · eligible outcome · exclusions · review and dispute rules
Buyer commitment: Written baseline, eligible outcome, data access, and shared execution responsibilities.
Measurement: The effect earned above the agreed baseline, not reported movement alone
Release condition: The baseline and eligible outcome are agreed in writing before any variable fee is discussed.

Work close to the evidence

Start with the decision, not a capabilities pitch.

Bring us the question your reporting can't settle. We'll tell you what evidence would actually move it, and whether we're the right people to help.

Work with us

Accountable judgment. Close to the evidence.

Every judgment has an accountable owner.

Name the decision before selecting the method.

Make the source record inspectable.

Test the explanation, not just the movement.

Assign the action and the rule that governs it.

Establish the truth

Set the next move

Operate the answer

Bounded diagnostic

Operating partnership

Outcome-linked structure

Start with the decision, not a capabilities pitch.