A ROAS in the twenties is either the best channel you have or the most expensive mistake you have not found yet. Origin decided to find out which.
Origin’s story matters right now because the weeks before Black Friday are when brands scale the channel that looks best, and the channel that looks best is often the one taking credit for sales it did not cause. Black Friday falls on November 27 this year, and most budget decisions for it are already being made on platform dashboards.
Origin is a useful case because it should have been the easiest brand in the world to fool itself. It has a loyal audience, a founder people follow, and products customers search for by name. When ads run, a lot of buyers who were already coming show up through the ad anyway. In the Origin case study published by Triple Whale, Justin Parker, Director of Ecommerce, described looking at Meta ROAS in the twenties and asking whether it was real. Most brands never ask that question at all.
Origin is a Farmington, Maine based brand that makes jeans, boots, jiu-jitsu gis and everyday apparel on a fully American supply chain, sewing in its own factories in Maine and North Carolina. Founder Pete Roberts started the company in 2011 with Brazilian jiu-jitsu uniforms, then expanded into denim and footwear; Elevation Outdoors reported that Origin now runs three factories across the two states. You can see the range at originusa.com.
That model shapes everything about how Origin markets. Making everything domestically means higher unit costs and a less forgiving supply chain than a brand sourcing overseas, so margins are tight and there is little room to spend on ads that do not pay back. Origin sells through Shopify, advertises on Meta, Google and Microsoft Ads, and runs email through Klaviyo, according to the case study.
Origin’s core decision was to stop adding up each platform’s revenue claims and start measuring every channel against one independent model. According to notes Triple Whale shared with eCommerce Fastlane ahead of a planned interview, the biggest problem Origin’s team faced was simple arithmetic: when they added together the revenue each platform claimed, it came to more than the business actually made. With thin margins, they could not plan spend around numbers that did not reconcile.
Justin came to that problem with a long view of where platform data goes wrong. Per the same notes, he ran marketplace selling across a dropship network with a large SKU count and plenty of data, and was an early beta tester for Google Shopping and Jet.com. People who have watched marketplaces report on themselves tend to be skeptical of any channel that grades its own homework.
The situation. Meta reported ROAS in the twenties. Klaviyo credited individual emails with large revenue figures. Each tool was confident and none of them agreed.
The approach. Origin adopted Triple Whale in March 2024, the case study says, for attribution that does not give every channel credit for every sale. The team then checked the claims one at a time. The clearest example in the case study is a single email that Klaviyo credited with $118K in revenue; Triple Whale’s model, which accounts for the other touchpoints in each customer’s path, put it at about $44K. The email still worked. It just did not work as well as the tool that sent it said it did.
The team also asked a question most brands skip: does ad spend drive the traffic that looks organic? The case study describes a correlation analysis, run through Triple Whale’s Moby assistant, between Origin’s blended spend on Meta, Google and Microsoft and its direct traffic over 60 days. It found a strong link, which gave the team grounds to defend increasing spend rather than cutting it.
The result. Meta held up. Per the notes Triple Whale shared, Meta’s return stayed strong and proportionally strong once measured independently, so the channel was earning its budget; what changed was that the team now knew where the efficiency threshold sat. The case study also reports that Origin’s BI team saved about 40% of its time, and that questions which took 30 minutes of manual analysis came down to about five.
The framework. Verify before you scale. For a strong brand, the risk is rarely that a channel is bad. It is that you cannot tell how good it is, so you cannot tell when you have pushed it too far.
Origin’s next step was letting AI take over the repetitive parts of paid media, starting with budgeting and bidding, while the team kept the decisions that need judgment. According to the notes Triple Whale shared, budget and bid management through Moby Media Buyer is the work Origin stopped doing by hand, and the team describes its ROAS this year as the best it has been.
The more interesting detail is what the AI surfaced that nobody asked for. Origin did not yet have formal incrementality testing, and per the same notes, Moby used outside signals to flag which campaigns were probably not incremental, with branded search as a clear example. That matches a pattern seen across the industry: branded search captures shoppers who would have found you anyway. Fastlane’s analysis of branded search cannibalization in Google Ads ROAS covers why that line item so often looks better than it is.
Origin is now moving toward formal measurement, with media mix modeling and incrementality testing next and work to understand what each platform and campaign type brings in. It has also started using Moby’s landing page tool, according to the notes.
Sonar Send is the other piece the case study puts numbers on. Sonar Send identifies more site visitors so Klaviyo can trigger abandonment flows it would otherwise miss. Origin reports $450K in incremental revenue within a year of setting it up, and 15% to 20% more events on top of the enrichment tools it already ran.
Origin’s approach required a dedicated ecommerce lead, a BI function, several overlapping data tools and enough monthly spend for attribution differences to be worth arguing over, which puts the full version out of reach for many smaller brands. The case study lists Elevar and Retention.com running alongside Sonar Send, so Origin layered tools rather than replacing them. That costs money and takes someone to keep the data clean.
It also took time. Origin started in March 2024, and the Sonar Send result is measured over the first year. The AI media buying came after the measurement work, not before it. A brand that skips straight to handing an AI its ad budgets, without first agreeing which numbers to trust, is automating the arithmetic problem Origin set out to fix.
The figures here are Origin’s and Triple Whale’s own, published by Triple Whale or shared in its interview notes, and have not been independently audited or tested with a holdout group. Triple Whale has also published the same Sonar Send testimonial at $420K rather than $450K, so treat the exact number as approximate.
Where it does not apply: a brand running one channel, or spending so little that a 10% reallocation would not matter, will get more from fixing its offer and creative than from a better attribution model. And a brand that does not make its own product may have more margin room to absorb a misread channel than Origin does.
This week, add up the revenue each of your ad and email platforms claims for last month and compare the total with what Shopify says you actually sold. If the platforms claim more than you made, you have Origin’s problem, and you should not scale any channel on its own reported ROAS until you know which claims overlap.
If you are doing under $5M, that comparison and a look at your branded search spend are enough for now. Pause or cap branded search for a short test after Black Friday, not during it, and watch whether total sales move.
If you are doing $5M to $50M, Origin’s sequence is the one to copy: get one independent attribution view, check your biggest channel and your biggest email against it, then hand repetitive bidding to software only once you agree on the numbers. Fastlane’s Triple Whale review covers pricing and plan limits if you are weighing that platform, and the explainer on incrementality in marketing covers the next step Origin is taking.
Disclosure: Triple Whale is a sponsor of the eCommerce Fastlane podcast, and this spotlight is part of that partnership.
A high Meta ROAS can be misleading because Meta counts sales from people who saw an ad, including people who were already going to buy. Brands with strong word of mouth or a well known founder see this most, since many of their customers arrive through an ad on the way to a purchase they had already decided on. Origin’s team saw Meta ROAS in the twenties and questioned it for that reason. The fix is not to distrust Meta, but to compare its claims with an independent attribution model and with total sales in Shopify before scaling spend.
You know your platforms are double counting when the revenue they each claim adds up to more than your store actually sold in the same period. Pull last month’s attributed revenue from Meta, Google, TikTok and your email tool, add it together, and compare it with Shopify’s total sales. Origin found this gap clearly in email: one send that Klaviyo credited with $118K looked closer to $44K once other touchpoints in the customer’s path were counted. A large gap means your channel ROAS figures should not be used to set budgets on their own.
Origin handed budgeting and bidding for paid media to Moby Media Buyer, according to notes Triple Whale shared, and kept the measurement and strategy decisions with its team. The AI also flagged campaigns that were probably not incremental, such as branded search, before Origin had formal incrementality testing in place. The order matters: Origin moved to independent attribution in March 2024 and automated bidding later, once the team trusted the numbers the automation would act on.
Independent attribution is usually worth it once you run two or more paid channels and the platforms start claiming the same sales. For a brand on one channel with a small budget, comparing platform claims with Shopify totals by hand once a month gives most of the insight for free. Origin’s setup, with a dedicated ecommerce lead, a BI team and several data tools, fits a brand with meaningful spend and thin margins, where a misread channel is expensive. Start with the manual comparison, then decide whether a platform would pay for itself.
You should not change your attribution setup or hand an AI your budgets for the first time in the final weeks before Black Friday. New tracking needs time to build history, and a peak weekend is the worst time to discover a misconfigured pixel. Use the time before BFCM to run the simple comparison of platform claims against Shopify totals, so you scale with your eyes open, and make structural changes in January when spend settles down.
