Funded Accounts vs First Trades: Prediction Market Tracking

Choose a product-specific event dictionary

Define registered, eligible, funded, first order and first completed trade separately. Name the product and system owner for each. The definition of funded must state whether it means accepted external funds, available balance or something else; promotional credit should not silently become customer funding.

Polymarket documents an order lifecycle with creation, matching and settlement stages. That is a concrete reason not to equate a submitted order with a completed trade. The actual platform’s accepted event and reporting controls must determine the metric.

Deduplicate people and event retries

A campaign can produce several orders or funding actions for one account. Count the first qualifying event once under a stable account boundary. Keep transaction or order volume as separate measures, and explain any reporting gap rather than filling it with a guessed user count.

Use authorised aggregate data. Public wallet activity does not give a campaign permission to identify a private person or combine unrelated records. Kalshi’s public affiliate notice explicitly warns against attempted re-identification and says dashboard estimates may change.

Separate referral credit from product activity

Kalshi’s referral FAQ explains that applying a code and meeting an offer’s requirements are separate steps. It also directs users to the specific offer terms rather than a universal reward rule. Do not treat a code attachment as proof of a qualified acquisition or import an offer amount from a different campaign.

For a brand reporting specification, retain referral source, qualifying activity, event time and status. Let the product owner resolve eligibility. A creator should receive an intelligible aggregate reconciliation, not private verification documents.

Test a fictional event sequence

Suppose 50 attributed eligible accounts receive accepted funding. Thirty submit an order. Twenty reach the agreed completed-trade event; several of the others still have resting orders. These invented figures illustrate state separation and are not a conversion benchmark.

The report can state 50 funded accounts and 20 first completed trades under the declared window. It cannot call all 30 order submitters active traders without changing the metric label. If a trade later fails under the product’s lifecycle, apply the agreed correction rather than preserving an inflated success count.

Make the report useful for a content decision

Compare equal elapsed follow-up and identify product failures separately from content delivery. If viewers understood the message but the destination was unavailable, the corrective action differs from a misleading brief. Do not pressure people to trade merely to improve the campaign funnel.

Keep final payment qualification in the applicable agreement. An analytics event, a referral reward and a creator commission may each have different rules. The useful output is an auditable aggregate event ledger, not a claim that creator exposure caused every observed trade.

A prediction-market creator brief needs different product and settlement questions from a sportsbook brief.