I Asked Practitioners One Question | Answers Told Me Everything | Satish Saka | MDU Engine

Added: 2026-05-20

[00:00:00]I asked the practitioners one question.

[00:00:02]The answers told me everything. I put a question to a group of people who work in growth and performance marketing. Not a complicated question, and the answers I got were so varied that I think they reveal something important about where the industry actually is right now. I am Alistair I research how decisions get made in data-driven environments. And last week I asked a simple question that gave me a complicated answer.

[00:00:30]The question was, how do you decide when a signal is reliable enough to act on?

[00:00:36]Not what tools do you use? Not what your target CPA is. Just how do you know when it's actually time to make the call?

[00:00:45]Some people said they wait for three consecutive days of positive trend, [music] 72 hours of consistent direction before they will consider moving.

[00:00:55]Others [snorts] apply a statistical significance threshold. They treat every optimization decisions like an experiment and won't act without a confidence interval.

[00:01:06]Some are honest, we mostly react to what we see. We have to build good instincts [music] over years and we trust them.

[00:01:13]Others have a fixed cadence. Weekly review and major calls only get made in that window regardless of what the daily data is doing.

[00:01:23]Others One person said, "Honestly, it depends on who is in the room. The most senior person's weight tends to be what we go with."

[00:01:32]And none of these approaches are wrong.

[00:01:35]Every single one of those people [music] is running campaigns that get results.

[00:01:40]But look at the spread.

[00:01:42]Three days of trend, statistical significance, weekly cadence, [music] pure instinct, and the senior person decides. These are not five different statics for the same underlying theory.

[00:01:55]These are five completely different theories about what a reliable signal even is.

[00:02:01]Sprout Social's 2025 Benchmark Report, based on analysis of 3 billion social messages, talks about how brands that consistently outperform benchmarks do so through systematic approaches. But the system for evaluating signal readiness, nobody seems to have agreed on one yet.

[00:02:23]I don't think there is one right answer in that list, but I do think the fact that experienced practitioners hold such different frameworks for the same fundamental question tells us the problem has not been properly solved.

[00:02:38]This is the space I work in. Not optimization itself, but the decision layer before it.

[00:02:44]>> [music] >> Which of those five sounds most like how your team does it? Or have you got a completely different approach?