The channel that was working looked exactly like the channel that wasn't.
Last-click reporting credited paid search for orders that started with a 30-second video. So the video budget got cut.
Snapshot
- Segment
- E-commerceDTC homeware
- Team size
- 75 content, 2 performance marketing
- Channels
- 5TikTok, Instagram, Pinterest, Shorts, Facebook
- Time to value
- 2 weeksStore and ad accounts connected in 3 days
Revenue attributed per post, across channels
01 — The situation
Before
This brand launches a product drop roughly every six weeks, and the whole business runs on that rhythm.
The shoot happens on a Monday: stills for the site, video for social, usually four SKUs. The content team cuts it Tuesday and Wednesday. The drop goes live Thursday at 9am, and the first 72 hours decide whether the drop was good.
Customers find these products the way people find homeware: a short video of the thing being used in a room that looks like theirs, watched on a phone, on a sofa, on a Tuesday evening. They do not buy then. They come back two days later, search the brand name, click the paid search ad sitting on it, and check out.
Every analytics surface the team had was reporting that last click. Paid search was credited with the order. Organic social appeared in the dashboard as 'direct' traffic or did not appear at all.
02 — The friction
What it was costing
The reporting was not merely incomplete. It was actively pointed at the wrong channel, and budget followed it.
Reconstructed later, roughly 61% of revenue with an organic-social origin was being recorded against something else. The content team was producing the material that started the majority of purchase journeys and could show no revenue for it. In the annual planning round, that translated into exactly what it always translates into: a proposal to shift budget from content production toward paid search, on the evidence available.
The team also could not answer a question their founder asked repeatedly and reasonably: which post made us the most money last year? Not which post got the most views — they had that — but which one produced orders. There was no way to get from a post to a purchase.
The last cost was speed. Occasionally a video obviously took off, and the sensible response is to put money behind it. Doing so meant a performance marketer rebuilding the creative inside the ad platform: re-uploading, re-cropping to the ad format, rewriting copy, setting up the campaign structure. That took about nine days from noticing to live, by which point the organic momentum it was meant to amplify had usually passed.
What this was costing
- Organic-social revenue credited elsewhere≈61%
- Orders matched to a source post39%
- Time to promote a winning organic post≈9 days
- Launch-day variant production1.5 days
03 — What changed
The reworked workflow
The store and the ad accounts were connected to the same place the posts go out from, which made the join possible at all.
- Revenue attribution from connected commerce sources
Posts are joined to orders, not to sessions
With the store connected as a source, revenue is attributed back to the post that started the journey rather than to the last link clicked before checkout. The founder's question became answerable in a filter: highest-revenue posts, last twelve months, ranked. The answer was not the post anybody had predicted.
- Cross-platform analytics
Discovery and conversion are read as one path
Reach, saves and watch time sit next to orders and revenue for the same piece of content. That is what made the pattern visible: the videos with the highest view counts and the videos with the highest revenue were substantially different sets, and the brand had been optimising for the first one.
- Promote top organic posts to paid
A winner gets budget in hours, not in a rebuild
A post that is already performing organically is promoted across the connected ad channels without reconstructing the creative in each ad platform. The nine-day rebuild became a decision made the same day the post starts moving — which is the only window in which amplifying it is worth anything.
- Upload once → platform variants
Drop-day content is produced from one master
Each SKU's video is uploaded once and comes back as vertical, square and feed variants with per-platform captions and titles. Launch-day production dropped from a day and a half of formatting to a couple of hours, which mattered most on the Thursdays when the shoot ran late.
Nothing about the content changed in the first month. The brand simply stopped being wrong about which content was working.
04 — Results
What moved, and by when
Orders matched to a source post
39%94%within 6 weeks
Time to promote a winning organic post
≈9 daysunder 48 hrsfrom week 3
Launch-day variant production
1.5 days≈2 hrssecond drop
Paid budget reallocated to proven organic creative
0%18%following quarter
Honest note on what drove this
Be careful with the first figure: it measures visibility, not growth. Those orders existed before — the brand was simply crediting them to the wrong channel. No revenue was created by connecting a data source. What changed is that the next budget decision was made against a roughly accurate picture instead of a systematically distorted one, and the reallocation is where any actual return comes from. Attribution across sessions and devices is never exact; treat 94% as 'matched with reasonable confidence', not as truth.
We were about to cut the budget for the only thing that was working, because the dashboard said the ad did it.
05 — What to take from this
Transferable lessons
Useful whether or not you ever open NOWScale — these are the parts that generalise.
Last-click reporting punishes discovery channels by design
Any channel that creates demand rather than capturing it will look weak under last-click. If your budget follows that model, it drains steadily toward the channels that sit closest to checkout.
Ask which post made money, not which post performed
View count and revenue rank different content. If you have never compared the two lists for your own catalogue, assume they diverge — they usually do, and the gap is where the budget should move.
The value of a winner decays in days
Amplifying a post that is already moving only works inside its window. If promoting takes over a week, you are consistently buying reach for content that has already peaked.
Better visibility is not more revenue
Fixing attribution changes what you know, not what you earned. The return arrives only when you actually reallocate on the new information — and that is a decision, not an outcome of connecting a data source.