How should we measure Snapchat results, and how much should we trust them?
Our answer
Treat Snapchat's reported results as a ceiling. Snap counts purchases up to 28 days after a click or 1 day after a view, so discount its return by Common Thread Collective's 0.30 times and confirm with totals and a holdout.
- Confidence
- Medium
- Strongest proof
- 8 official rules
- 2 datasets, 3 practitioners, 6 opinions or claims
- Sources
- 9 sources
- 19 statements
- S
New here? Take the 1-minute tour
Why this is the answer
3 reasons and our judgmentSnapchat's dashboard overstates results. Across 299 ecommerce brands in Q1 2026, it reported a 6.18 times return against 1.85 times after an incrementality factor (Common Thread Collective, ecommerce ad agency). Marin Istvanic saw three new channels each report a 7 ROAS while revenue stayed flat.
Snap's counting rules lean generous. By default it counts conversions up to 28 days after a click and 1 day after a view, reports them on the impression day and adds modeled conversions. It expects gaps against Google Analytics.
Snapchat can still pay even when its dashboard overstates. Measured (incrementality testing firm) found from its geo-calibrated panel that Snapchat indexes at 1.9 times the average channel brands add after Google and Meta.
Our judgment
Until your own test, multiply Snapchat's reported return by Common Thread Collective's 0.30 times before comparing it with other channels, and expect that factor to move, since it rests on one quarter's benchmark and Measured's geo-calibrated panel ranks Snapchat well above the average new channel. Read Snapchat's purchases on a click-only window next to the default before comparing with Meta or Google Analytics, because the default credits views and Snap itself says to match windows across platforms. Use Snap's A/B tests and Brand Lift for choices inside Snapchat, and a geo holdout for whether Snapchat adds sales. Its A/B test picks the lowest cost per goal on Snap's own attribution, and Brand Lift measures survey answers.
Our judgment, built from Common Thread Collective, Measured, Snap, Cometly and Balistro: our reading of these sources, not a finding from any one of them.
What the sources say
19 statements from 9 sourcesSnap recommends its default 28-day click, 1-day view attribution window. Since November 2024 it sends 5-second video views to app measurement partners as clicks, which can raise Snap's click-attributed conversions.
“We recommend advertisers utilize a 28 day click and 1 day view attribution window (28/1), which is the default on our platform.”
Since August 22, 2024, Snapchat Ads Manager reports conversions by impression time by default, using the 28-1 attribution window. Snap suggests matching the view your third-party reporting uses.
“Starting on August 22, 2024 Ads Manager will default to reporting based on Impression Time instead of Conversion Time using the 28-1 attribution window.”
Snap says its Estimated Conversions are modeled from engagement, advertiser profile and aggregated conversion data, blended into standard Ads Manager columns on a 28-1 window by default, and used in bidding.
“Estimated Conversions are a form of statistical modeling that's used in our attribution methodology”
Snap says its numbers will differ from Google Analytics, which counts same-session conversions while the Snap Pixel counts cross-device ones up to 28 days after a view; match UTMs, windows and time zones.
“Attribution Windows should match across platforms”
Snap says its split tests pick the ad set with the lowest cost per goal as winner, only after each arm has spent at least $10. It compares ad sets on Snap's attribution and holds no one out.
“A winner will only be determined once each arm of the Split Test has reached at least $10 spent.”
Snap's Brand Lift splits an audience at random into exposed and unexposed groups and compares their answers to a short in-app survey; it measures perception, not sales.
“brand awareness and other key perception metrics among people who have seen your advertising and those who have not”
Snap offers a 28-day click, 1-day view optimization window, and a 7-day click window for advertisers with 10+ click conversions in 7 days. It recommends matching delivery and attribution windows.
“In general, we recommend matching your delivery window as closely as possible to your attribution window.”
Marin Istvanic says one client added Pinterest, TikTok and Snapchat, each reporting 7 ROAS, but total revenue did not change. If MER falls with the new spend while revenue stays flat, the channel is just taking credit.
“if you add a new channel and MER drops proportionally to the new spend while total revenue stays flat, that channel is stealing attribution.”
PPC Land says Snap's claim of 19.3% higher incremental return than blended social comes from studies Snap paid for or chose, gives no sample or date, and its footnote cites 12.9%.
“come from studies Snap paid for or chose to cite”
Snap cites a Measured study finding Snapchat delivered about 19.3% higher incremental ROAS than brands' social advertising overall. It gives no sample size, date or link for the study.
“Snapchat delivered approximately 19.3% higher incremental return on ad spend than the blended incremental return from their social advertising overall.”
Measured's panel of 269 expansion-channel programs put Snapchat at 1.9x the panel's incrementality average, with conversion buys far ahead of awareness buys. It says to size tests with a power analysis.
“Size the test with a power analysis, not a round number.”
Snap says Northbeam clients cut new customer cost 20% on Snapchat last Black Friday and Cyber Monday, and its Third-Party Web Optimization beta drove 37% lower CPP. These are Snap's figures for its own product.
“Last Black Friday / Cyber Monday, Northbeam clients cut the cost of winning a new customer by 20% on Snapchat, while every other major platform got more expensive.”
Snap's Value Optimization for ROAS is in beta. Snap says that in a test with Google Analytics, its Third-Party Web Optimization gave 23% lower third-party cost per purchase and 37% higher ROAS.
“advertisers saw 23% lower third-party cost-per-purchase and 37% higher third-party return on ad spend.”
PPC Land reports Snap opened Unified Attribution to app advertisers using AppsFlyer or Adjust. Snap's beta results each come from one advertiser and show a changed counting rule, not proven lift.
“the numbers describe attributed outcomes under a changed counting rule, not a demonstrated lift in business volume.”
Balistro, an Indian agency selling paid media, says Snapchat suits brands targeting under-25 buyers at about 15 to 20% of paid social, judged on incremental CAC and LTV from a geo or holdout test.
“It works best as 15-20% of a paid-social portfolio, measured on incremental CAC and LTV”
Snap's lead ads now connect to HubSpot, sending lead fields in and conversion events back through the Conversions API. Snap claims lead CTR rose 77% and cost per lead fell 26% year over year.
“Snapchat lead form fields mapped into HubSpot properties”
Admiral Media, a Snapchat agency, says testing Unified Attribution against Snap Outcomes for one dating app gave 38% more Android installs and 11% lower cost per install, read in AppsFlyer, not a holdout.
“We tested Snapchat Unified Attribution while it was still in beta.”
Cometly, which sells attribution software, says Snapchat defaults to 28-day click and 1-day view attribution, which credits view-only conversions, so standardize windows across platforms before comparing.
“Snapchat Ads Manager defaults to a 28-day click and 1-day view attribution window.”
In CTC's Q1 2026 benchmark of 299 DTC brands, Snapchat's platform ROAS of 6.18x fell to 1.85x once adjusted for incrementality, and Pinterest's 6.57x to 2.76x. Each took under 0.5% of total spend.
In the full answer
For readers who sign in- What to do this week3 steps
- The numbers behind it3 figures from cited sources
- Where each source standsA chart placing each source on the question
- The strongest disagreementThe best case against the answer, from a cited source
- How sure we areWhat its confidence rests on, and what would change the answer