The Ecomm Analyst

A client showed me a slide last quarter that had their company-wide LTV:CAC ratio at 3.2. Healthy. They were thinking about increasing paid spend and wanted my read.

I asked them to break it apart by acquisition channel. They sent me the file two days later. Meta was at 4.1. Google search was at 5.8. Influencer was at 1.4. TikTok was at 0.9.

The 3.2 number was real. It was also useless. It was averaging two channels that were quietly losing money against two that were printing it. The leadership team thought they had a healthy business. They had two healthy businesses and two unprofitable ones, blended together into a number that felt fine.

The fix isn’t complicated, but it is unglamorous. You have to compute LTV by acquisition channel, which means tagging customers cleanly enough at the moment of acquisition that you can follow them out 12, 24, 36 months. Most brands cannot do this in their main commerce stack. The data is too fragmented across platforms.

I do this in ThoughtMetric because it’s the only place I can see acquisition channel and lifetime value in the same view without rebuilding the analysis every quarter. The mechanics matter less than the principle, though. If you have a way to do it, do it. If you don’t, build the way.

A few things that surprised me the first few times I ran this analysis on client accounts.

Influencer almost always looks worse than the brand thinks. Customers acquired through influencer drops tend to be one-and-done. They came for a specific creator’s recommendation, not for the brand. LTV is often 30 to 50% lower than customers acquired through other channels, which means the CAC needs to be correspondingly lower for the channel to pencil out, and it usually isn’t.

Branded search almost always looks better than the brand thinks. The CAC is low because customers were already searching for you. The LTV is high because they’re coming in with high intent. The catch is that branded search isn’t really an acquisition channel. It’s a closing channel for demand created somewhere else, and if you cut that somewhere else, branded search dries up too.

Email and SMS, when treated as acquisition channels (meaning the customer’s first purchase came through a flow), tend to have the highest LTV ratios in the dataset. This isn’t because email is magic. It’s because those customers had already opted in, which selects for higher intent.

The right way to look at LTV:CAC is as a portfolio. The blended number tells you whether the portfolio is healthy. The channel-level numbers tell you what to actually do. If you only have the blend, you have a thermometer. You don’t have a steering wheel.