Kochava launched AIM Essentials on November 11, 2026. The company says the product uses one to two years of daily event and media spend data. Before handing over that much history, check one week of exports for problems that could carry through to the rest of the dataset.
Why a one-week audit matters
A short review window is not training data. It is a sanity check. The goal is to confirm that the KPI definition matches the business question, timestamps are aligned to the same timezone, and channel spend reports are being interpreted consistently across sources.
This kind of pre-modeling review is also a good time to separate a missing export from a genuine zero. Those can look the same at first glance, but they mean very different things for modeling and interpretation.
What to verify in the sample week
- Outcome: name the key performance indicator (KPI), such as completed orders, and use the same definition across exports.
- Timezone consistency: confirm that the event day and spend day are using the same cutoff.
- Spend: reconcile the channel total against the matching report in the same currency and date range. Document differences in fees, credits, or reporting cutoffs instead of treating a close match as proof.
- Missing export versus true zero: a blank file, failed job, or delayed sync should not be treated as zero spend or zero conversions.
- Promotion flagging: note any sale, coupon, pricing change, or other unusual promotion that may explain a spike or dip.
A simple hypothetical discrepancy
Suppose an ad platform reports a charge on Sunday at 8:30 p.m. New York time. In November, the same moment is Monday at 1:30 a.m. UTC. If your warehouse groups events by UTC but the spend export groups by New York time, the records can land on different reporting days even when both systems captured the charge.
Record both cutoffs and compare the same reporting window before calling the gap missing spend. If you only have daily aggregates, ask for an export aligned to the chosen time zone instead of shifting every row by one day.
General pre-modeling context
Google Meridian’s pre-modeling guidance also emphasizes collecting clean data before model fitting, including consistent time periods, clearly defined outcomes, and reliable media spend inputs. That is general context rather than a Kochava-specific requirement, but the underlying principle is the same: quality checks come before modeling.
Google Meridian pre-modeling data collection guidance
The illustration below uses mock figures and does not show the Kochava product or actual campaign results.
