Enter password
Clicks vs. Survey · 12 Months

Which channels over-report — and which are invisible?

Every Fairing survey response carries both the customer's answer to "What led you to purchase today?" and the actual ad click (UTM) on the same order. 65,818 paired first orders, July 2025 → July 2026, let us compare what the tracker saw against what the customer said — order by order, no statistical modeling.

GENERATED Jul 13, 2026 · FAIRING Q2864 (FIRST ORDERS) × NORTHBEAM CLICKS-ONLY, CASH · SPEND & CLICK NCs: NORTHBEAM · NC TOTALS: DAILY SPEND TRACKER

65,818
Paired orders
click + survey, same order
2.7×
Search over-credited
clicks claim 14.6%, customers say 6.3%
0.46×
YouTube under-credited
clicks see less than half its role
0.01×
TikTok invisible
1.9% of customers, ~0% of clicks
70.9%
Meta self-agreement
Meta clickers who credit Meta

Credit Ratio — click-attributed customers ÷ survey-implied customers

1.0× = the click model and customers agree. Above 1.0 = channel harvests credit created elsewhere. Below 1.0 = channel creates demand the clicks never see. 12-month totals.

The Matrix — what the click said vs. what the customer said

Rows = the tracked click source on the order. Columns = the customer's survey answer. Row-normalized. Diagonal cells (click and answer agree) are the channel's self-consistency; big off-diagonal cells are borrowed credit.
Click source ↓ / Customer says →Meta AdYouTubeWord of MouthOnline SearchTikTokPodcastSomeone I FollowAn ArticleVideo GameOther
Meta clickn=20,90270.9%2.2%5.4%2.8%2.3%0.7%2.5%1.6%0.3%7.6%
Google clickn=9,19728.7%16.2%18.6%9.8%2.4%3.8%2.4%1.7%0.5%9.5%
YouTube clickn=3,1937.6%70.8%3.2%2.7%1.2%2.5%0.9%1.3%0.5%4.8%
Email/SMS clickn=4,47141.4%11.0%11.2%6.1%2.2%1.4%1.8%2.3%0.7%11.9%
Affiliate clickn=2,43326.1%6.8%15.9%22.8%1.4%2.1%1.8%5.3%0.5%10.9%
AppLovin (game ads) clickn=84623.2%7.6%9.2%3.5%2.5%0.8%1.8%3.0%29.3%13.8%
Other paid/UTMn=1,58128.3%13.2%15.1%8.5%1.4%3.9%2.3%2.8%0.6%15.6%
No UTM (organic/direct)n=23,18938.1%15.6%13.4%6.8%1.6%3.4%2.8%1.8%0.8%9.5%

CAC, Two Lenses — 12 Months

Click CAC = spend ÷ Northbeam clicks-only customers (the number in every report today). Survey CAC = spend ÷ survey-implied customers (survey share × Shopify-truth new customers). Caution: survey answers credit the channel including its organic presence — for TikTok especially, the survey mostly reflects organic content, not the small ad spend.
ChannelSpendClick NCsClick CACSurvey NCsSurvey CACΔ
Meta$10,919,04772,106$15186,835$126-17%
YouTube$1,984,00811,851$16726,030$76-54%
Search$1,693,89133,106$5112,066$140+175%
AppLovin (games)$452,5142,528$1791,696$267+49%
TikTok$52,97246$1,1523,581$15-99%
CAC by Channel — Two Lines, Monthly

Northbeam click CAC vs. Fairing survey-implied CAC

Red = the CAC in today's reports (spend ÷ clicks-only customers). Green = spend ÷ (Fairing share of that month's answers × Shopify-truth new customers). Where green sits far below red (YouTube), clicks overstate the cost; where green sits above red (Search, AppLovin), the reported CAC flatters the channel. Gap = survey off Sep 2025. TikTok omitted — its 46 annual clicks make a click-CAC line meaningless.
Meta
YouTube
Search
AppLovin (games)
Stability & Spend Response — Monthly
Credit ratio by month — is the bias stable? (log scale; 1.0 = agreement)
YouTube — monthly spend ($K) vs. survey share
Meta — monthly spend ($K) vs. survey share
Search — monthly spend ($K) vs. survey share

Scatter panels: each dot is a month. A rising pattern = survey share responds to spend (the channel's claimed influence is buyable). Seasonality confounds Q4 dots — treat as directional, not causal.

Read
Search is a harvester, not a creator — budget it like one. Clicks credit Search with 14.6% of customers at a $51 CAC; customers themselves say 6.3%, implying ~$140. Only 1 in 10 Google-click purchasers says search led them — 29% say Meta, 19% word of mouth, 16% YouTube. Brand search is capturing demand other channels (and your brand) created. Keep it for defense; stop celebrating its CAC.
YouTube's real CAC is closer to $76 than $167. Customers credit YouTube with 2.2× what clicks capture, and YouTube clickers self-confirm at 70.8% — exactly as reliable as Meta's 70.9%, so the survey signal is trustworthy. Combined with its Q4 surge to 22% of discovery, YouTube looks like the most under-priced line in the budget. The spend↔share scatter backs it: months with more YouTube spend show higher claimed share.
Even Meta is under-credited (0.83×). 38% of purchasers with no tracked click at all still credit Meta — its ads create demand that converts later, organically. Meta's true reach is bigger than its already-big click numbers; its problem remains price per customer, not presence.
TikTok: the survey sees ~3,600 customers; clicks see 46. With only $53K of annual spend, that's mostly organic TikTok content doing the work — a channel influencing ~2% of customers with almost no investment or measurement. This is the clearest green-field signal in the data.
AppLovin is mildly over-credited (1.5×) — and the matrix shows why. Only 29% of game-ad clickers credit games; 23% credit Meta. Roughly half of AppLovin's clicked conversions look like borrowed demand. Its survey-implied CAC ($267) is worse than its click CAC ($179) — the opposite direction of every creator channel. Worth a holdout test before scaling further.

Method & caveats: Credit ratio = channel's share of Northbeam clicks-only customers ÷ channel's share of survey answers (share-of-share, so blended totals cancel out). Survey CAC uses Daily Spend Tracker D2C new customers as the base. Surveys measure claimed influence (memory, multi-touch) — a channel can influence without being "the" cause; ratios are evidence of direction and rough magnitude, not precise incrementality. Only a holdout/geo test proves incrementality. Survey off Aug 18 – Oct 12, 2025. YouTube cluster includes Agentio (creator sponsorships); Search = Google + Google CA + Microsoft; Meta includes Canada. TikTok survey answers reflect organic + paid.