Skip to main content

Analytics

Bytes records activity when somebody uses a tracked Link.

Analytics connects that activity to Links, channel and UTM context, Activity Groups and Offline Engagement where the available data supports that attribution.

Useful Analytics on every plan​

Free is not an upgrade-only Analytics screen.

The current Free Analytics history is 7 days, with useful headline performance and supported basic context.

The Dashboard also uses a 7-day history window on Free.

Open a Link's Analytics page and switch from Overview to Clicks to inspect the individual interaction events behind the aggregate metrics.

Every plan with Link Analytics access can see:

  • when the event happened;
  • whether Bytes classified it as Human or Bot.

Free keeps this view deliberately shallow inside the existing 7-day Analytics history.

Starter and above can also see the current stored click context where available:

  • country;
  • device class;
  • browser;
  • operating system;
  • source, including explicit Direct and Unknown states.

Starter and above can filter the click list by those available dimensions. The existing Reporting period control applies to both Overview and Clicks, so there is not a separate click-history date model.

A click is an interaction event, not a known person. Bytes does not expose raw IP addresses, IP hashes, raw user-agent strings or raw referrer URLs in this view.

Current geography is country-level only. Region and city are not inferred where Bytes has not stored that information.

Starter​

Starter expands Analytics with a current history window of up to 90 days and additional breakdowns.

Current Starter capabilities include supported:

  • performance breakdowns
  • audience timing/context
  • custom date ranges
  • current-view export

The exact sections shown depend on the data available in the workspace.

Pro​

Pro supports up to 365 days of Analytics history.

Current Pro capabilities build on Starter with features including:

  • previous-period comparison
  • deterministic Needs Attention checks
  • online/offline mix where the data supports it
  • full-detail export
  • AI Recommendations when platform Intelligence is enabled and available

Needs Attention is not a fake AI label. The first implementation uses deterministic checks to identify practical measurement weaknesses and next actions.

AI Recommendations and the Watchlist​

AI Recommendations analyse bounded workspace activity and suggest practical next checks. On the Dashboard, the recommendation period follows the selected Dashboard period so the recommendation context and the chart refer to the same recent activity window.

Where the workspace has comparison entitlement and the previous period is still within Analytics retention, the recommendation input also includes an immediately preceding period of the same length. This gives Intelligence a real basis for identifying material changes rather than asking it to infer change from a current-period snapshot.

The provider input is intentionally minimised. It uses performance totals, ranked performance counts and controlled categories needed for the analysis. It does not send destination URLs, raw visitor interaction rows, IP hashes, browser user-agent strings, or customer-authored Link, Offline Engagement and asset names, descriptions, slugs or UTM values.

When a recommendation set is generated by AI, Bytes labels it AI Recommendations. A cached recommendation set is also identified as cached rather than presented as a new generation.

The Watchlist is different. It is deterministic and remains useful when AI Recommendations are disabled, unavailable, not included in the workspace entitlement, temporarily unable to refresh, or do not yet have enough activity to analyse.

Bytes does not silently present deterministic Watchlist checks as AI output.

AI Recommendations are currently included on Pro and Business when platform Intelligence is enabled. Enterprise access is subject to the workspace's commercial entitlement. Promotional entitlements can also grant access where explicitly configured.

Recommendation actions are validated by Bytes before they are shown. The AI provider does not receive permission to write directly to the Bytes database or execute customer changes.

Business and Enterprise​

Business and Enterprise currently use the same 365-day standard Analytics history window.

Business adds deeper reporting capabilities such as:

  • saved export definitions
  • run-now saved exports
  • advanced export capability
  • Offline Engagement asset comparison where supported

Scheduled report packs remain separate future work until they are visible in the product.

Date ranges​

Bytes validates ranges against the workspace plan and available history.

A custom range cannot expose data outside the workspace's permitted retention window.

Attribution should stay honest​

Bytes does not turn incomplete signals into false certainty.

Examples:

  • direct or unknown referrers are labelled rather than invented
  • shared physical Links can create shared attribution
  • unique visitor figures are estimates from privacy-conscious identifiers
  • small samples should not be treated as strong evidence
  • device/manufacturer detail is only shown to the level the stored signal supports

See Understand your results for interpretation guidance.

Last updated on