Why Triangulation Beats Any Single Survey Score

The engagement score is the most trusted number in HR — and one of the least reliable. Not because engagement doesn't matter, but because a single number from a single source isn't evidence. It's a data point. There's a difference, and it matters enormously when you're trying to act on it.

The single-source problem

Survey scores are subject to response bias (who responds and why), timing bias (a bad week inflates negative responses), social desirability effects (people say what they think is expected), and Hawthorne effects (being observed changes behavior). Any one of these can move your score by several points without any actual change in employee experience.

When leadership sees a 3-point drop in engagement, they call an all-hands. When it goes up 4 points, they celebrate. But without triangulation, neither signal means much. You don't know if the change is real.

What triangulation means in practice

Triangulation means using three independent sources of signal and only trusting a finding when at least two of them agree. In EX work, those three sources are typically:

1. Quantitative survey data (engagement pulses, lifecycle surveys, manager effectiveness ratings)

2. Qualitative signal (journey interviews, focus groups, open-text responses analyzed for theme)

3. Behavioral traces (turnover data, internal mobility rates, absence patterns, promotion velocity by demographic)

When all three point the same direction, you have a finding worth acting on. When only one does, you have a hypothesis worth investigating.

The practical threshold

In our Diagnostic work, we apply a simple rule: we don't put a finding in a client-facing report until at least two independent sources confirm it. This slows the work down slightly and prevents us from chasing noise. It also means that when we say 'there is an experience gap in onboarding for junior hires,' we can show the survey data that shows it, the interview themes that describe it, and the 18-month attrition rate that proves the cost of it.

That kind of evidence is defensible in front of a board. A single survey score is not.

Start with what you have

Most organizations already have two of the three sources. They have engagement survey data. They have attrition data. What they often lack is the qualitative layer — the human voices that explain what the numbers are showing. Adding even a small number of structured journey interviews (15-20 per persona) transforms a pattern into a story.

Stories move people. Numbers frame the problem. The combination produces change.