> ## Documentation Index > Fetch the complete documentation index at: https://docs.sourcemedium.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Data and attribution health > Check whether your tables are fresh and how much of your acquisition traffic carries valid UTMs, straight from Slack Two questions tell you whether an analysis can be trusted. Is the data current? Is marketing traffic being tracked? The Slack AI Analyst answers both on demand. | Check | Ask | What it answers | | ------------------ | ------------------------------- | ---------------------------------------------------------------------------------- | | Data Health | "How is my data health?" | Are my tables fresh and populated? Which domains are ready to analyze? | | Attribution Health | "How is my attribution health?" | What share of my acquisition orders carry a valid UTM? Where is tracking breaking? | These are separate problems. A table can be fresh while attribution is poor, and tracking can be fine while a table is stale. Check Data Health first. If the data is fresh but the numbers still look wrong, check Attribution Health. You can also click **Data Availability Report** in the opening message of a new DM thread. It lists your tables, their status, and your available stores. ## Data Health Data Health reads the freshness metadata for every SourceMedium table in your warehouse. It does not query the tables themselves. | What it reports | What it means | | --------------- | --------------------------------------------------------------------------------- | | Stale tables | Tables with no fresh data in the last 14 days | | Empty tables | Tables with no data at all | | Ready domains | Which analytical areas are usable: orders, customers, ads, attribution, and so on | When every key table is fresh, the report says core analytics are ready. When a key table is stale or empty, it names the table and the analyses that are affected. The data health report also includes a short attribution health summary, so one question covers both. ### When data updates Most tables refresh daily. Data is fresh through the previous day in your reporting timezone. Today's data is incomplete, so do not use it for analysis or reconciliation. Some sources lag longer. See [Data freshness](/docs/help-center/core-concepts/data-transformation/data-freshness). ### Why 14 days Core tables such as orders, customers, sessions, and ad spend should refresh every day. The 14-day threshold exists for secondary tables that may not see daily activity. Examples are a new subscription program, a refunds table with low volume, or an integration that only fires on specific events. If a core table is stale, that is almost always a pipeline issue. If a secondary table is stale, ask whether you would expect activity there. ### If data is degraded 1. Ask "Which tables are stale?" to see whether one source or several are affected. 2. Scope your analysis to dates the data covers. If orders stopped syncing on the 15th, do not analyze the 16th onward. 3. If staleness lasts more than a day or two, [open a support request](/docs/product-tours/get-support). It usually means a connection needs attention. Example questions: ``` How is my data health? Which tables are stale? What data do I have available? When was my orders data last updated? ``` ## Attribution Health Attribution Health runs a live audit of your online DTC orders from the last 30 days. Marketplace orders from Amazon, TikTok Shop, and similar platforms are excluded. Those orders are attributed to the marketplace by definition, so there is no UTM to capture. ### The status The headline status is the share of acquisition orders that carry a valid UTM. A valid UTM excludes `(direct)`, `(none)`, self-referrals, and sessions broken by a payment provider redirect. | Valid UTM coverage | Status | | ------------------ | -------- | | 80% or more | Healthy | | 60% to 79% | Warning | | Below 60% | Critical | ### What the report includes | Line | What it tells you | | ----------------------------- | ---------------------------------------------------------------------------------------------------- | | Valid UTM coverage | Share of acquisition orders with a usable source and medium | | Invalid or direct/none bucket | Share of acquisition orders with no usable source | | Payment-provider breakage | Orders whose session was reset by a redirect through shop.app, PayPal, Klarna, or a similar provider | | Self-referral breakage | Orders where your own site was recorded as the source | | Landing page coverage | Share of online DTC orders with a recorded landing page | | Funnel session coverage | Share of online DTC orders with session-level data | | MTA journey coverage | Share of online DTC orders with a multi-touch journey | The report also lists your top source and medium pairs, landing pages, referrer domains, discount codes, and zero-party survey responses, plus recommended actions based on what it found. The report has no trend. To see whether coverage is improving, compare it with a report you ran earlier. ### Common causes of poor coverage * **Missing UTM parameters.** Campaign links without UTMs arrive as direct traffic. * **Payment provider redirects.** Checkout through shop.app, PayPal, Klarna, and similar providers can start a new session and drop the original source. * **Self-referrals.** Your own domain gets credited as the source after a redirect. * **Cookie and tracking blockers.** Browser privacy features prevent the session from being recorded. ### How to improve it UTM setup, checkout attribute capture, post-purchase surveys, and more. Standardize UTM tagging across campaigns. Causes of missing attribution and how to fix them. Capture UTMs at checkout so a redirect cannot drop them. Example questions: ``` How is my attribution health? Is my UTM coverage good? Why is so much traffic direct/none? What percentage of orders are unattributed? ``` When numbers still do not match after both checks, see [Troubleshooting](/docs/ai-analyst/troubleshooting#results-do-not-match-expectations).