The questions we hear most.

An agency replaces the operator. Allocation replaces the operating system. We run the funnel architecture, the audience exclusions, the budget math, and the measurement. Your team (or your agency) keeps owning creative, brand voice, stage-specific messaging, and the offer calendar.

The other difference is the contract: an agency typically charges a percentage of spend, with no efficiency floor. Allocation is a flat $7,500/month with the fee at risk if we miss the platform floor.

A marketing mix model tells you, in retrospect, what worked across channels. Allocation does the prospective version: it forecasts each platform's Low / Mid / High band for the next 30 days, reallocates spend daily against that band, and tells you which days are worth pressing on before you spend.

No. The funnel restructuring works on your existing creative. Over time, the stage-and-tier structure makes it easier to test creative cleanly because each ad set has one job, but no creative change is required to start.

A qualified cart is one built with verified product, price, and identity signals, not just a fired "add to cart" event from a botted or duplicate impression. A qualified checkout is the same standard at the next stage. These qualified events are what gate the funnel transitions and what feed back to the Top so cold prospecting compounds over time.

The current engagement covers Meta and Google. The same model framework extends to TikTok and Amazon; those typically come in as Phase-2 add-ons after the 90-day POC.

Week zero. The Edgemesh tag deploys on your domain, the Meta Business Manager API connection is established, the 90-day historical backfill loads, and the baseline attribution dashboard goes live. By Week 1, Edgemesh is managing 25% of Meta spend; the remaining 75% runs as a control split.

Each month, Allocation produces a Low / Mid / High benchmark band per managed platform. The "Low" is the floor. If actual results miss that floor for a platform in a given month, 100% of that month's fees for that platform are refunded. The benchmark is shared with your team before the month starts, so there's no after-the-fact moving of goalposts.

Allocation reads from server-side, first-party event capture via the Edgemesh tag, not from Meta's reported numbers. That makes the dependent variable in the model independent of Meta's modeling. Signal loss makes platform-reported numbers worse over time; it doesn't degrade Allocation's data.

Daily ROAS varies meaningfully even when budget and creative are held constant. That's market seasonality, and it's the part you can actually model. Allocation forecasts daily because the band moves daily; less than daily, an operator can keep up, and more than daily, platform noise drowns the signal. Aggregate accuracy is a consequence of getting the days right.

The model is fit weekly on your account and on the broader Edgemesh corpus, so it tracks platform-level changes as they happen. The structural insight that the platform optimizes for its own earnings inside a band has been stable across every algorithm change we've seen.

Yes, and we encourage it. Allocation handles the structural and quantitative side; your operators keep handling creative, brand voice, offer strategy, and stage-specific messaging. The 25% / 50% / 75% ramp during the POC is specifically designed to let both run side by side.

A 12-month contract at the same flat $7,500/month, with the guarantee in force every month. Most brands expand the portfolio in Year 1, adding TikTok or Amazon, and use Q4 as the first full peak season on Allocation.