Why Your Data Recommendation Didn't Land (And How to Fix It)

You can be prepared, practiced, and genuinely knowledgeable, and still lose a room — not because your material was bad, but because you made the audience work too hard to extract what matters. Explaining is not the same as communicating. Whether in a meeting room, a 1-on-1, or Slack, these principles hold.

Data work is already magic to most stakeholders. The pipeline runs, the dashboard updates, and the room nods along with roughly the same understanding a theatre audience has of the trapdoor.

Researchers ran eye-tracking studies on magic performances and found that some observers looked directly at the secret move and still missed it. The perceptual system has to allocate attention and compare the right representations. Without that, you can stare at the answer and register nothing.

The data consulting parallel is obvious. You can put the correct recommendation on slide three, say it clearly, and still watch an executive go with their gut feeling. Correctness is not salience.

The Two Failure Modes

Most technically inexperienced communicators fail in one of two ways:

The Oversharing Engineer

Floods the room with caveats and detail until the decision disappears. They think they're doing something noble: "I owe it to the work to be thorough."

More explanation rarely saves you though. Once you push past what the room can process, people stop reasoning through the substance and start reaching for shortcuts. The more you know, the worse you get at seeing where the gaps are.

The Bunker Engineer

Strips out too much, hides uncertainty, and creates false simplicity. Their lie sounds noble too: "I'm protecting them from complexity."

It's easy to become cynical with people and enamoured with technology and not wanting to speak up and deal with the fuzzy reality that lives in a business.

What do both of these personalities have in common? Their communication style serves themselves, not other people.

The Semmelweis Problem

Semmelweis is the cleanest case of someone being technically right but completely losing the argument in the delivery. In the 1840s, he worked at Vienna General Hospital and discovered that doctors were carrying death on their hands — moving from autopsies straight into the maternity ward with no hand hygiene.

He required everyone to wash their hands with chlorinated lime. Mortality dropped fast. On the merits, this should have been a triumph.

But psychologically, it was a grenade. He implied the doctors — people who dedicated their lives to saving lives — were the ones killing people. As the stakes rose, he escalated. He denounced opponents publicly. He called one prominent critic a "medical Nero."

He was optimizing for vindication, not adoption. And the more he pushed, the more his audience's incentives shifted from "Is he right?" to "We cannot let him win."

Once you make the conversation about someone's identity, you stop arguing with their model. You start arguing with their self-protection. And self-protection almost always wins.

The 3-Layer Communication Model

No matter how clearly you explain a technical system, the underlying complexity will leak through — usually at the worst moment. You cannot stop the leaks. You can decide what surfaces first.

Layer 1: The Decision

This is what 90% of stakeholders need:

The decision or recommendation in one sentence
Why it matters now
Expected impact as a direction and magnitude range
The top two tradeoffs
The next step: who does what by when

Layer 2: The Known Risks

This is where you proactively plug the leaks you can predict:

Assumptions that must stay true for this to work
Risks and failure modes you will monitor
Constraints from legal, budget, or politics
Alternatives you considered and why they lost

Layer 3: The Appendix

Method, data quality notes, diagrams, and deeper model details. This is where engineer-brain can breathe, and where the audience goes when they need to verify. Most people will never leave Layer 1. That is fine.

Translation Frameworks

Pick the one that fits the room, not the one that sounds smartest:

Framework	When to Use	Structure
BLUF (Bottom Line Up Front)	Busy, senior, action-oriented audiences	Recommendation → Evidence → Method
SCQA	Resistant audiences or bad news	Situation → Complication → Question → Answer
What / So What / Now What	Quick translation test for any message	Fact → Meaning → Next action
1-3-1	Trade-off decisions	1 problem → 3 genuine options → 1 recommendation

Action titles — Slide titles should state the conclusion, not the topic. "Churn analysis" is a topic. "Churn concentrates in the first 90 days" is a usable sentence.

All of these frameworks share one principle: think from the bottom up, present from the top down. You did the analysis in one order. The audience needs to receive it in the reverse.