---
title: "Metrics that misled: the flat traffic that wasn’t"
url: "https://toddpaulbrownjr.com/writing/metrics-that-misled/"
author: "Todd Paul Brown Jr."
description: "A client's flat traffic hid growth and a conversion problem. \"Why now?\" is the question that separates a metric from a story."
kind: "article"
updated: "2026-09-26T02:50:35+00:00"
---

# Metrics that misled: the flat traffic that wasn’t

A B2B restaurant-operations software company was leaving around 2024. Organic traffic looked flat on the dashboard, and flat was enough to cancel.

I pulled the work chat thread with my business partner. The message I sent: "ooooohhh.... those numbers are coming from data falling off in search console...."

Search Console's default lookback window is shorter than most people remember. Traffic that falls outside that window just disappears from the chart. The client's flat line wasn't flat at all — it was a reporting artifact.

I ran longer-range analytics instead. From 2022 to 2024, organic traffic had grown 9.5%, from about 48,000 to 53,000 (measured). Website form submissions went from 335 to 564 in the same period, a 68% increase (measured).

My partner replied that demo requests — the outcome metric the client actually cared about — had stayed flat. So the metric misled in both directions: a tool artifact hid the growth, and the growth didn't move the number that mattered.

## "Why now?"

In July 2025 I interviewed for a paid-ads director role. The assessment scenario was a real-estate-investor client threatening to cancel after "8 days without a lead." This was an assessment exercise, not a real client, but the problem was worth solving.

I checked the history first. Conversions, cost per lead, and overall performance had been pretty steady for the last year. The question I wrote in my prep notes: "The thing I want to identify is 'why now?' If the account has been steady for such a long time, what is different right now that makes the client feel more like pulling the plug on their ads?"

The 8-day gap was within normal variance. Qualified conversions averaged roughly five and a half days apart, and performance had been steady for about five quarters (measured from the scenario data).

If the data showed the cancellation was more emotional than logical, I needed to play to that emotional decision-making — not argue with the feeling using numbers the client had already discounted.

## Start from revenue

My approach is to always start with revenue and work back from there, with impressions being the very last metric I would look at. And you have to jump in and play around with it to see what anomalies stand out.

A flat total can hide a real change underneath. The restaurant client's traffic window artifact hid growth, and the growth hid a conversion problem. A steady account that suddenly feels broken is usually steady in the wrong thing, or the person on the other end hit a threshold that has nothing to do with the chart. Ask "why now?" before you argue with the emotion.

Read the composition before the total. The metric is just a number until you know what moved and what didn't.