I used to use Google Trends to forecast Chipotle traffic for the purposes of trading the quarter. Given Chipotle’s recent demand woes, it was helpful to identify weakness in same store sales YoY growth.

Google Trends worked — until it didn’t.

Going into Q4 2025, simply using the term “Chipotle” was sufficient to forecast same store sale (SSS) traffic. The data suggested a YoY increase in SSS. Given the negativity going into the quarter, I thought that the stock would pop.

Chipotle reported a 3.2% drop in traffic.

So what did I miss?

The mix shift to digital orders, which skewed the data.

Google Trends “Chipotle app” searches :

• Oct: +27% YoY

• Nov: +50%

• Dec: +84%

My model assumed that Google Trends could tell us about overall demand. But the data likely told us about online digital orders. The company did report a 200 YoY bps in digital orders. in Q4. But perhaps adoption of the app is accelerating.

The crisis of affordability also emerged meaningfully in Q4. There has been a significant rise in Google search terms like “cheap lunch”.

For forecasting Q1 ‘26, I created a multi-variable regression analysis using Google Trends data now including the terms “chipotle app”, “cheap lunch”, and “chipotle menu”. I created some time lags for the data as well. The R² was over 0.90. Admittedly, there are only 9 data points.

Google still has close to 90% market share of search, so the issue of LLMs taking market share won’t be an issue, at least not today.

For Q1 2026, the regression suggests traffic is flat YoY.

Chipotle’s 2026 guidance was full-year SSS at flat.

So the lack of confidence in the model, coupled with an output that validates full year guidance, suggests no trade.

At some point there will be a favorable YoY inflection when SSS increases. That’s when the stock likely pops. Unclear when this happens, and to what extent the company has to maintain or lower prices to get back to the affordable value proposition the company used to have.

Investors I know avoid this type of trade entirely because they assume others have better data, like credit card transactions. I used to think Google Trends could provide that edge. The takeaway for me is that relying on a single keyword or a model is not sufficient. Particularly because the underlying behavior can shift.

I’m not saying Google Trends is useless. But I am saying that I’m no longer comfortable using it as a primary input for trading this or any other name.

Disclaimer: This is for entertainment purposes only, do your own research.