Methods
How Cernia works on an account.
The same procedures, run the same way, on every account: read everything live, reconcile before concluding, draft the move, prove the result on your own numbers. Your approver says yes before anything changes.
House rules
The rules every number follows.
Growth claims are cheap. These six rules decide what Cernia will and will not print, before any method is applied.
Attribution is the floor.
Your attribution tool's click credit is each channel's floor. Platform-reported conversions are a health signal, not revenue.
Cost comes from the platform.
Spend is read from the ad platform, never from an attribution tool's cost import. No cost, no cost-per-anything printed.
Clicks, not views, for social.
View-through credit is left out unless you ask for it, and then it is labelled.
Fresh campaigns on conversion date.
A recently raised campaign is graded on when orders happened, not when clicks did.
Three lenses, reconciled.
First-click, linear and platform numbers are compared before any conclusion is drawn, and the disagreement is shown to you.
One alarm is never a conclusion.
A drop is checked against processing lag, hourly completeness and an independent control before anyone is told traffic fell.
Procedures
Ten things that happen on every account.
Read everything live
Every platform you spend and measure on is read at the moment a question is asked or a procedure runs: search, shopping, social, analytics, attribution, call tracking, lifecycle and commerce. Nothing is copied into a warehouse, so there is no stale export and no second version of the truth.
Reconcile before concluding
Platforms grade their own homework. Cernia puts first-click, linear and platform-reported numbers side by side, applies the house rules above, names the number it used and shows where the lenses disagree.
Audit on a schedule
Standing audits run on every account: search terms that never convert, keywords bidding against each other, winners starved by a budget cap, placements nobody chose, tracking that quietly broke. Each finding arrives as a plan with its fix, ready to approve.
Move the money inside guardrails
Each month the budget is drafted toward what is bringing in new customers, inside guardrails you set: protected brand campaigns, a lifetime-value gate, change caps and a paid-search floor. Spend-neutral by default, until a result earns the next dollar.
Build the ads from what already sells
Ad copy is written from your best sellers, your search terms and your reviews. Variants are built and launched paused; you approve; the winners are scaled. A platform you have never run is connected within a day and launched with researched structure, targeting and copy rather than an empty account with a budget attached.
Run fair experiments
Campaign experiments and geo holdouts are registered with their stop rules before they start and sized so the read is real. Winners roll out. Losers are logged with the reason, so the same mistake is not bought twice.
Measure what clicks cannot see
A Bayesian media-mix model on your own weekly data measures the channels attribution undercounts. It only adds to click credit, never lowers it, and channels it cannot read are flagged rather than quoted.
Write back under a contract
Every change is a plan with an ID: dry run, approve by ID, apply, read back, roll back. New objects start paused. Nothing is deleted. Budgets, bids and targeting are not touched until you enable it, and only a named approver in your organization can say yes.
Keep the numbers
Every dashboard figure and recurring-report figure is kept as a series at daily, weekly, monthly, quarterly and yearly grain. When a platform restates history, the old point stays and the change is marked, so leadership's comparison does not quietly move.
Remember why
Every decision, rule and lesson is logged, so each month starts from what the last one learned and a new question is answered with the account's history in hand.
Start
What the first month returns.
Day one
Every connection verified, a data-health report for each platform, and the first audit of each account: the waste it found and the fixes, ready to approve.
Week one
The cross-platform paid-media review with its spend-neutral moves, tracking and attribution repaired, your weekly report scripted to your format, and the first new ad variants built and launched paused.
Month one
Budget moves calibrated to your data, reports and dashboards live, the first experiments running with their stop rules, your team asking Cernia questions in Slack and getting answers with receipts, and a first media-mix model where the history allows it.
Questions
What people ask before the first call.
- What do we have to provide?
Read access to each platform through its own sign-in, your definitions (what counts as a new customer, your attribution model, revenue basis, goals), and one person on your team who approves changes. Cernia brings the reading, the drafts and the checks; your team keeps the decisions.
- Does this replace our attribution tool or our dashboards?
No. Cernia reads the attribution tool you already have and treats its click credit as the floor. It adds the hours nobody on the team has: reconciling every platform, deciding what to do, drafting it, and proving the result.
- How are changes made to our accounts?
Only through a plan with an ID that your approver says yes to. The plan names every field with its current and intended value, a rollback plan exists before the first change, and nothing is ever deleted.
- How fast does a new platform connect?
Any platform with an API typically connects within a day. The committed list is on the home page; others are added as work items with the same contract.
- What does a report look like?
Headings, tables and charts set the same way on screen, in the PDF and in the workbook, with every figure carrying its source, its window and the checks it passed. The example report on this site shows the form on simulated data.
- Which model reads our data?
The one you choose. Run on Cernia's account or register your own Anthropic or OpenAI key. Scripts summarise platform rows before a model reads them, customer contact data is never read, and nothing you connect trains anything.
See it on your accounts.
Write with your site and the platforms you spend on. The first call opens with what we found.