15 Attribution Views Across 7 Dimensions: What DTC Analytics Should Actually Look Like

Key Takeaways

Most DTC operators at the $500K-$2M revenue range know their top-line number.
A 2024 Rockerbox study found that only 22% of DTC brands have multi-touch attribution in place, and fewer than 10% can attribute revenue at the SKU-variant level.
A typical business at $1M revenue can answer "what is our refund rate?" with "about 5-8%, I think."
If you are making CPL, ROAS, or LTV decisions based on platform-reported aggregates, you are flying blind.

The Setup

Most DTC operators at the $500K-$2M revenue range know their top-line number. Some know it by product. Almost none can tell you revenue by product by affiliate by campaign by month, with initial/rebill/refund splits, per SKU variant, across two currencies.

That gap is not a reporting inconvenience. It is a decision-making blind spot. When you cannot attribute revenue to its source at the unit level, every decision about where to spend, what to scale, and what to kill is based on aggregates. And aggregates lie. A 5:1 blended ROAS looks healthy until you discover that one campaign is running at 12:1 while three others are running at 1.5:1 -- and you are feeding the losers equally.

The DTC operators who scale past $1M profitably are not the ones with the best products. They are the ones with the deepest visibility into where every dollar comes from and what it costs to acquire.

What the Data Shows

A 2024 Rockerbox study found that only 22% of DTC brands have multi-touch attribution in place, and fewer than 10% can attribute revenue at the SKU-variant level. Triple Whale's 2023 DTC benchmark report showed that brands using granular attribution made 37% better ROAS decisions compared to those using platform-reported metrics alone. Nielsen's 2024 marketing mix study confirmed that brands with attribution depth across 5+ dimensions outperform those with basic analytics by 2-3x in marketing efficiency.

Stealth Labz built a revenue attribution system that produces 15 structured data exports across 7 attribution dimensions, tracking $1.08M in gross revenue across 38 products, 14 affiliate sources, 38+ campaigns, and 28 months of transaction data. Every dollar is traceable from the traffic source that generated it to the product it purchased to the month it occurred in.

The seven dimensions: product (38 families, 100+ SKU variants), affiliate (14 tracked IDs), campaign (38+ funnels), time (28 months), transaction type (initial/rebill/refund), currency (ZAR and USD), and payout (affiliate cost per offer per source).

Any combination can be queried. Revenue by affiliate by month. Refund rate by product by campaign. Initial vs. rebill split by SKU variant by affiliate. The system answers questions most operators at this revenue level cannot even ask.

How It Works

A typical business at $1M revenue can answer "what is our refund rate?" with "about 5-8%, I think."

This system answers: "6.0% on PRD-01 ($32,535 / $542,356), 11.2% on PRD-02 ($16,138 / $144,932), 8.6% on PRD-03 ($9,446 / $110,355). Weighted portfolio average: 7.2%. AFF-01-sourced refund rate is higher than STL-sourced. Here is the monthly trend."

That depth informed every major decision in the portfolio. The PRD-01 decline was visible in real-time -- monthly revenue by affiliate by product showed exactly when AFF-01 traffic was dropping and what it meant for the product arc. The affiliate dependency was quantifiable because every dollar was tagged to its source -- when AFF-01 accounted for 96.9% of revenue in February 2024, the system flagged the concentration risk in data, not in hindsight. The product launch discipline worked because the operator could see within 30-60 days whether a new product had traction, measured by initial orders by product by source by month.

The 15 structured exports cover the full attribution stack: revenue by product (USD), detailed product time series, Konnektive-native product attribution, raw and USD-normalized campaign revenue, campaign order counts, monthly revenue by affiliate, affiliate revenue with type splits, affiliate-by-SKU counts, affiliate-by-offer conversions, affiliate-by-offer revenue, affiliate-by-offer payouts, transactions by affiliate-by-product, and initial order counts by source by product by month.

This is not a dashboard. It is a unit economics engine built on Konnektive CRM transaction exports.

What This Means for DTC Operators

If you are making CPL, ROAS, or LTV decisions based on platform-reported aggregates, you are flying blind. Platform ROAS and actual ROAS diverge by 20-40% in most DTC verticals because platforms over-attribute and under-report refunds.

The minimum viable attribution stack for a scaling DTC operation needs four things: revenue by product by source by month, refund rate by product by source, initial vs. rebill split by product, and cost of traffic by source by offer. If you cannot produce those four views from your current infrastructure, every budget allocation decision is a guess dressed up as analysis.

Fifteen views across seven dimensions is not overkill at $1M revenue. It is what DTC analytics should actually look like when you are making decisions about which products to scale, which traffic sources to invest in, and which campaigns to kill. The depth does not slow you down -- it removes the guesswork that wastes your budget.

References

Rockerbox (2024). "DTC Attribution Report." Multi-touch and SKU-level attribution adoption rates.
Triple Whale (2023). "DTC Benchmark Report." ROAS decision quality with granular attribution.
Nielsen (2024). "Marketing Mix Study." Attribution depth and marketing efficiency correlation.
Keating, M.G. (2026). "Case Study: The Revenue Attribution System." Stealth Labz. Read case study
Keating, M.G. (2026). "The Compounding Execution Method: Complete Technical Documentation." Stealth Labz. Browse papers