MTA and MMM, on every conversion you actually had.
Multi-touch attribution and marketing-mix modeling that work because the orders are not a sample. They are every order, joined to every touch, on the same ledger. Plus a campaign manager, A/B tests, ad integrations, and a Marketing Moments calendar agents read.
The shape of marketing in CCEN.
Marketing is the attribution and campaign app. Touches stream in from your ad platforms, your site, your email tool, and your SMS tool, and join to Orders by exact id and fuzzy match. Multi-touch attribution recomputes on every order land. Marketing-mix modeling (MMM) tunes to your seasonality, your channels, and your Moments calendar. Campaigns plan against the same calendar. Consent-mode and iOS 14.5 cookie deprecation are handled at the touch layer, not as a marketing problem.
Not a sample. Every order.
Most attribution products work on a sampled subset, because they are running on data they piped in from other tools. CCEN's attribution runs on every order in the ledger, joined to every touch in the touch ledger. The numbers reflect what happened, not what the model guessed from a fraction.
Touches arrive from Meta, Google, TikTok, Reddit, native ad networks, your email tool, your SMS tool, and your own site click stream. Each touch carries its source, its medium, its creative, its placement, and the user identifier the platform reported. CCEN resolves user to customer, then attaches the touch to the order.
MTA models are pluggable: last-click, first-click, position-based, time-decay, data-driven. Each model produces credit per touch per order. The dashboard shows them side by side. The CFO can pick the one she trusts. The agency can pick the one they trust. The argument is on the model, not the data.
Mix modeling on your seasonality.
Marketing-mix modeling (MMM) needs a long history of spend, conversions, and outside factors. CCEN has them, because every order is in the ledger and every spend record is in the campaign manager. The MMM tunes weekly. The model accounts for diminishing returns, the lag between impression and purchase, and how channels reinforce each other, all estimated against your actual data.
Your sales calendar feeds the model so a Memorial Day spike is recognized, not treated as random. A spike during your sale is a known holiday with a known lift profile. Models attribute spike correctly. Spend outside known sale windows gets evaluated on its own merits.
Outputs are budgets and uplift estimates, exportable to your media plan. The agency sees the constraints. The CFO sees the assumptions. Incrementality testing runs in the same surface, so 'lift the model claims' has a holdout test you can show your investor.
Plan, run, learn, on one surface.
The campaign manager is one surface for every channel: paid social, paid search, email, SMS, organic posts, retail collateral. Each campaign carries dates, budget, creative variants, and the segments it targets. Spend records and conversion records bind to the campaign by id.
A/B tests are first-class on the listing, the email, and the campaign. Variant assignment is deterministic per customer. Lift is measured against the holdout, not against last week. Significance thresholds are configurable. Results land on the campaign timeline next to the spend.
When a campaign closes, the system writes a structured retrospective: spend, conversions, attributed revenue per model, learnings flagged. Marketing reviews are a list, not a thread.
Three things Triple Whale and Northbeam, running on a sample, will not do.
Attribution on every order, not a sample
Every order, every touch, on the same ledger. The CFO and the agency look at the same numbers. The argument is on the model, not the data.
MMM that knows your sale calendar
Memorial Day is not unexplained variance. Your sales calendar feeds the model. The forecast is honest about which lifts came from spend and which came from holidays.
Incrementality tests in the same surface
A/B tests with deterministic variant assignment, holdout-based lift, and significance thresholds. The lift the model claims has a test you can show your investor.
Attribution, planning, testing, on one ledger.
MTA, multi-model
Last-click, first-click, position-based, time-decay, data-driven. Side-by-side. Auditable inputs.
MMM, weekly tune
Diminishing returns, lag-to-purchase, channel interactions. Sales calendar fed in. Outputs are budgets.
Marketing Moments
Holidays, sale events, launches, campaign windows. Read by replenishment, marketing, and forecasting agents.
Campaign manager
Paid social, paid search, email, SMS, organic, retail collateral, all on one surface with budget and dates.
A/B and incrementality tests
Listing, email, campaign. Holdout-based lift. Significance thresholds. Sharable retros on close.
Ad platform integrations
Meta, Google, TikTok, Reddit, native. Push spend, pull conversions, sync custom audiences.
Email and SMS sync
Klaviyo, Postscript, Attentive, native CCEN flows. Touches join to orders. Segments flow both ways.
Custom audiences
Push CCEN segments to Meta, Google, TikTok. Refresh on schedule. Suppressions respected.
Consent-mode handled
iOS 14.5, EU consent, cookie deprecation, all handled at the touch layer. Marketing teams stop wrangling pixels.
Pick a model, see the credit per touch.
The same week's orders, costed across the four attribution models most operators argue about. Toggle to see how each model splits credit across touches. The data underneath does not move. Only the model does.
100 percent of credit goes to the last touch before purchase. Easy to read, but biases toward the channels that close (search, retargeting), and undercounts the channels that introduce.
- Meta credited: 18%
- Google credited: 41%
- Email credited: 31%
- Direct credited: 10%
100 percent of credit goes to the first touch. Useful for measuring how a channel introduces customers. Undercounts the channels that close.
- Meta credited: 38%
- Google credited: 22%
- Email credited: 17%
- Direct credited: 23%
Credit decays from the order back through every touch with a configurable half-life. The closer the touch to purchase, the more credit it gets, but earlier touches are not zero.
- Meta credited: 28%
- Google credited: 33%
- Email credited: 26%
- Direct credited: 13%
Credit splits learned from your own conversion patterns. The model finds the touches that actually move conversion, not the ones that happen to fire last.
- Meta credited: 31%
- Google credited: 28%
- Email credited: 23%
- Direct credited: 18%
The ad platforms, email tools, and analytics rails your team runs.
First-party integrations to the ad platforms, email tools, and analytics rails the marketing team already runs. Custom audiences sync both ways.
“We were paying for an attribution tool that ran on a 12 percent sample. Switched to CCEN's MTA on all orders and discovered our 'best' Meta campaign was actually being credited for organic traffic we'd already paid for. Cut spend, lift went up.”
Attribution that earns the budget.
Connect your ad platforms and your email tool, and see MTA and MMM running on the orders you actually shipped.