Solo Stove

CHALLENGE

Measure the true, incremental ROAS of every marketing channel

Solo Stove’s marketing data lived in a dozen platforms, each with its own login, its own definition of a conversion, and its own claim on the credit. Last-click attribution over-credited the channels that touched a customer last and under-credited everything that built demand earlier. Paid media data often arrived only as monthly aggregates, Amazon data came from multiple partners and needed heavy validation, and no single view tied upper-funnel spend to downstream sales.
The stakes were real. Leadership was weighing a budget shift toward influencer marketing, and the team needed data-driven scenarios before moving money. Liz’s wish list featured a few must-haves above all else:

1. One governed source of truth unifying every marketing, ecommerce, and advertising system
2. A measurement approach that estimates true incremental contribution by channel, not just last touch
3. Daily-grain data, available in time for in-season decisions
4. A clean split of upper-funnel awareness from lower-funnel performance, so ROI stays honest

SOLUTION

Turning fragmented spend into one model of what actually works

CorralData connected Solo Stove’s full stack into one platform delivered through a Snowflake share, then built the reporting and modeling layer on top. Salesforce Commerce Cloud, Amazon Seller Central, Amazon DSP, Google Ads, Meta Ads, Microsoft Advertising, Google Analytics 4 across four regional properties, and the Impact affiliate platform were unified and normalized, including currency normalization for non-US marketplaces and SKU-level cost and fee lookups so revenue could be read against margin. A custom session-level GA4 attribution model improved channel accuracy beyond the platform default.

Benefit #1: A region-aware marketing mix model

The centerpiece is a marketing mix model that answers the incrementality question directly. CorralData applies adstock decay so spend is credited across the days it keeps working, fits a multivariate regression on a trailing 365 days of data separately for each region, and produces a marginal ROAS for every channel. Because every channel enters the regression at once, the model apportions credit between channels that spend in sync rather than letting each claim the same sale. Every step from raw spend to dashboard number is visible in SQL.

Benefit #2: MMM next to last-click, side by side

CorralData puts the model’s incremental ROAS next to in-platform and last-click numbers on the same board, so the team sees exactly how much of a platform’s claimed credit is genuinely incremental. Coefficients re-fit automatically as the data refreshes, and a drift view tracks how each channel’s contribution shifts over time.

Benefit #3: Plain-English answers through AskCorral

Solo Stove’s team asks questions of their unified data through AskCorral, CorralData’s AI natural language query interface. Project instructions are set so media questions always run through the marketing mix model, keeping every answer consistent no matter who is asking. The same AI layer powers an automated executive summary and scheduled alerts that reach leadership on a regular cadence.

Benefit #4: Governed views for a full-funnel and a performance audience

Separate full-funnel and performance-only workspaces with role-based access mean the CMO sees the complete picture including upper funnel, while channel owners see the performance views relevant to them. Upper-funnel sources such as Meta awareness, YouTube, CTV, and Amazon DSP are tagged and excluded from performance views by default, which keeps ROI reporting clean.

Benefit #5: Daily-grain data, in time for the season

Paid media reporting moved from monthly aggregates to daily detail, and connectors are monitored for reliability. When the Impact affiliate connector dropped credentials, it was opened, fixed, and resolved within days, and dashboard and reporting change requests are turned around in hours rather than weeks.

RESULTS

From guesswork to budgets managed against incremental ROAS

Solo Stove now runs marketing measurement on one unified platform and treats the marketing mix model as the default lens for media decisions. Instead of debating which platform’s self-reported numbers to believe, the team manages spend against incremental ROAS. Liz uses CorralData to adjust marketing budgets and monitor the effect on performance, and the proposed influencer budget shift can now be pressure-tested as a data-driven scenario rather than a hunch.
Blended advertising efficiency has climbed as Solo Stove concentrates budget on the channels the model proves are genuinely incremental rather than the ones that simply touched the last click. The marketing mix model also exposed how wide the gap between channels really is, with the strongest channel delivering roughly three times the incremental ROAS of the weakest, giving the team the confidence to shift spend toward what actually works. And it serves as a control: when a Meta account issue distorted in-platform reporting and shifted apparent credit between channels, the MMM made the distortion visible and kept the team from misreading it as a real change in performance.