The most common complaint from hiring managers about analytical staff is not technical weakness. It is that findings arrive in a form nobody can use. Fixing that is a learnable skill with a small number of specific techniques.
The rule that governs everything
Lead with the decision, not the method. Nobody in a business meeting wants to know that you ran a two-sample t-test. They want to know what to do and how confident they should be. Method belongs in an appendix, available if challenged.
Plain-English translations that actually work
Statistical significance → "The difference is bigger than the normal week-to-week wobble, so it is unlikely to be a fluke."
Confidence interval → "Our best estimate is 6%, and the true figure is very probably between 3% and 9%."
p-value → Do not use it in the room. If pressed: "there is only a small chance we would see a gap this big if nothing had really changed."
Standard deviation → "How much this number normally bounces around."
Correlation without causation → "These move together, but we have not shown one causes the other - stores that got extra staff were already the busy ones."
Sample size → "This is based on 240 customers out of 9,000, so it is indicative rather than definitive."
Regression → "A formula that estimates X from Y, based on past patterns."
Statistically underpowered → "We do not have enough data yet to tell a real effect from noise."
The four-sentence structure
Open every findings conversation with this:
- What we found: "Version B increases conversion."
- How big: "By about 0.6 percentage points - roughly £40,000 a year."
- How sure: "Confident it is real; the plausible range is 0.1 to 1.1 points."
- What we recommend: "Roll it out, and re-check after a month."
Then stop and take questions. The most common presentation failure is spending nine minutes on method and running out of time before the recommendation.
Techniques that raise trust rather than lower it
- Volunteer your caveats first. "This excludes lapsed customers, so treat the retention figure as optimistic." Stakeholders who discover a limitation themselves stop trusting you; ones you told stop worrying.
- Use money and time units. "£40,000 a year" lands; "0.6 percentage points" does not.
- One chart, one message. If a chart needs a paragraph of explanation, it is the wrong chart.
- Give the range verbally, every time. It becomes normal, and then nobody expects false precision from you.
What to avoid
Jargon used as authority, refusing to simplify because "it is more complicated than that" (it always is - simplify anyway and note the caveat), and hiding uncertainty to sound more useful. That last one destroys credibility permanently the first time a confident number turns out wrong.
Why this is a career skill, not a soft skill
Analysts who can do this get invited to the decisions. Those who cannot produce correct work that gathers dust. The pay difference between the two is substantial and grows with seniority - see presenting findings that get acted on.
The Statistics for Business course at London School of Business UK treats communication as part of the discipline, not an afterthought. Enquire today.