Platform-reported numbers sit beside what Kompound measured — in their own column, never added together. Five Attribution models, your choice of window, and a ROAS adjusted for what those Customers go on to spend.
| Channel | Ad Spend | Orders | Measured ROAS | Reported | LTV-adj 90d | Net Profit |
|---|---|---|---|---|---|---|
| Meta Ads | $18,402 | 221 | 1.6× | 2.4× | 1.9× | $3,104 |
| Google Ads | $9,880 | 148 | 2.1× | 2.6× | 2.3× | $4,212 |
| TikTok Ads | $5,120 | 61 | 1.4× | 3.1× | 1.5× | −$318 |
| Klaviyo | $412 | 109 | 34.9× | 41.0× | 36.2× | $6,880 |
| Organic & Direct | — | 49 | — | — | — | $2,410 |
| All Channels · Coverage 91% | $33,814 | 588 | 3.5× | not summed | 3.9× | $10,318 |
Every Order matched to the visits that earned it, on the model you choose — shown beside what each platform claimed, so the gap is visible.
| Creative | Spend | LTV-adj ROAS | Status |
|---|---|---|---|
| UGC Video #3Meta · 728d live | $4,744 | 2.02× | Hot |
| TikTok UGC #2TikTok · 1039d live | $2,442 | 1.88× | Hot |
| Google Video UGC #2Google · 764d live | $3,186 | 2.79× | Fresh |
| Testimonial Static #4Meta · 851d live | $7,180 | 0.75× | Cooling |
| Offer Carousel #1TikTok · paused | $5,037 | 0.31× | Paused |
| Fatigue Alerts | 7 creatives cooling | Refresh | |
| Source | Sessions | Rev / Session | Orders |
|---|---|---|---|
| Google Organic | 18,402 | $1.42 | 188 |
| ChatGPTAI referral | 1,204 | $3.18 | 41 |
| ClaudeAI referral | 486 | $2.94 | 15 |
| Meta Paid | 24,110 | $0.88 | 221 |
| Direct | 9,880 | $1.06 | 96 |
| Site speed | LCP 2.8s on mobile | −$1,240/mo | |
Seven Reports that answer the same question from different angles: is this spend paying for itself?
Five attribution models beside what each platform claims. Switch models and watch which Channels keep their ranking — those are the ones you can trust.
Drill from Channel to campaign to ad set to the single ad, and explore every UTM by source, medium, campaign, content and term.
Every Klaviyo and Postscript send with the Orders, Profit and 90-day LTV it produced — flows and campaigns kept apart.
See what your winning ads have in common, catch creative that is tiring before the spend is wasted, and learn which ads bring Customers who buy again.
Which sources send buyers, which search queries earn Revenue, how AI assistants refer shoppers, and whether site speed is costing you sales.
Six real storefront stages from the Kompound Pixel — session, product view, add to cart, checkout, payment, purchase — split by device and Channel.
Ask buyers how they heard about you on the confirmation page, then read their answers beside click-based attribution.
Build consistent tags at the source, so Campaigns never fills up with three spellings of the same campaign.
Ask “which Channel brings Customers who come back?” and get the Cohort Breakdown that answers it.
"Is TikTok working?" is four Reports and a spreadsheet. Ask Kompass reads the same Attribution data and answers in one line — with the Report behind it.
How Ask Kompass Works →Meta counts a conversion if the buyer saw an Ad within its own window, and every platform does this for itself, so the totals overlap. Kompound matches each Order to one Channel on your chosen model. Both numbers are shown — never added.
Last Non-Direct is the sensible default. The more useful habit is switching models and watching which Channels hold their ranking — the stable ones are the ones you can trust.
Revenue from the Customers a Channel acquired, followed for 90 days, divided by the Spend that acquired them. A Channel that looks weak on day one often looks different at 90 days.
Only for the Funnel and Session-level numbers in Website Intelligence. Attribution, Campaigns and Email work from your Store and Ad account data alone.
Measured ROAS beside what the platform claims, Channel by Channel, from tomorrow morning.