Frequently Asked Questions
Does Klar send conversions back to the ad platforms?+
No. Conversion enrichment - returning attributed, server-side conversions to the platforms so their algorithms optimize against them - is not available in Klar. Its first-party script matches the same person across sessions and devices, and the result stays inside Klar's reports. Admetrics streams enriched conversions to six destinations, Meta, Google, TikTok, Snapchat, Outbrain and Taboola, as part of core tracking, after first-party server-side collection recovers the 20-30% of conversions browser pixels drop.
Can Klar move my budget for me?+
No. Klar's MMM, incrementality testing and targets show where budget should go; the change itself happens by hand in each ad account. Admetrics closes that step with the Budget Allocator, which turns attribution and MMM into a cross-platform allocation and executes it inside explicit objectives and constraints, with approvals where you want them, a full record of what moved and why, and rollback on every action.
Does Klar track profit, or only ROAS?+
Klar tracks profit. COGS, fulfillment, transaction and overhead costs run down to contribution margin, and its attribution reports carry margin, so channel decisions can be made on contribution rather than platform ROAS.
How many attribution models does Klar have?+
Five static models plus a data-driven one: first click, last click, linear, U-shape and unique, then a model that reweights each touchpoint by intent, order and time lag. Admetrics ships nine out of the box and lets you write your own rules for splitting credit, which matters when the way your business actually earns an order does not match any standard shape. Model count is the smaller question anyway - what feeds the model decides whether the output is worth acting on.
Does Klar have an AI assistant or an MCP server?+
Neither ships as a product. There is no in-app assistant to ask a question of, and no Klar MCP server - what exists is a documented API for imports and attribution reports, plus a published guide for building an MCP server yourself. Admetrics ships both: Ava answers profit and performance questions from your own data in plain language, and a native MCP and Data API give Claude or ChatGPT governed access to the same numbers and controls.



