Start here9 min read

Why Meta, Google, your CRM and Sissel show different numbers

Learn what each system counts, why totals disagree and how to reconcile a difference without forcing false parity.

The systems answer different questions

A difference is not automatically an error. Each system sees a different part of the journey and applies its own definitions. Before comparing totals, write down the exact business question you expect each number to answer.

SystemWhat it primarily knowsQuestion it can answer
Ad platformActivity and conversions reported to that platformWhat received credit under this platform’s rules?
Website analyticsSessions and actions observed on the websiteWhat happened during tracked website visits?
CRMPeople, stages, deals and recorded valuesWhat did sales accept, progress or close?
SisselConsented visits linked to CRM outcomesWhat source received credit under this named model and evidence?

Common reasons totals differ

  • Different outcome definitions

    One system may count form submissions while another counts unique qualified people or won deals.

  • Different attribution rules

    A platform can credit itself under its own click or view rules while Sissel applies your selected source-based model.

  • Different windows and dates

    Reports may use the click date, outcome date, upload date or a different timezone and lookback.

  • Identity and consent gaps

    A CRM record may exist even when no consented browser journey can be linked to it.

  • Duplicates and revisions

    A CRM can contain duplicate people, repeated webhooks or a deal whose value changed after its first creation.

  • Processing delays

    CRM syncs, ad-platform imports and delivery acknowledgements do not necessarily arrive at the same moment.

  • Missing campaign information

    An untagged link or stripped click identifier can leave a real visit without a usable campaign source.

A practical reconciliation sequence

  1. 1
    Name the outcome

    Compare the same event, such as qualified lead—not “conversions” in one system and all leads in another.

  2. 2
    Match the reporting window

    Use the same start, end and timezone. Confirm whether the report groups by visit date or outcome date.

  3. 3
    Match the attribution definition

    Document model, lookback and whether views, direct visits or cross-device assumptions are included.

  4. 4
    Compare a sample of records

    Open several individual journeys on both sides instead of reasoning only from aggregate totals.

  5. 5
    Classify each difference

    Separate expected definition differences from missing collection, identity, mapping or delivery evidence.

  6. 6
    Fix the data path, then rerun

    Correct the source issue and save a fresh named comparison. Do not rewrite the definition merely to make totals equal.

When a difference is useful

A stable, explained difference can be healthy. For example, the CRM may show more qualified leads than Sissel because some came from referrals, manual entry or journeys without usable website evidence. The correct result is to label those outcomes honestly, not assign them to a campaign without proof.