What is backlink velocity, and how do you track it in Metabase?

Backlink velocity is the net rate of referring-domain acquisition — domains gained minus domains lost per week or month. It's the rate-of-change companion to referring domains: the total measures where the link profile stands, velocity measures whether link earning is actually outrunning the constant churn of links that die.

What backlink velocity measures

Link profiles are leaky buckets: pages get deleted, links get edited away, domains expire. Velocity measures the fill rate against the leak — the net flow of unique referring domains per period. That framing is what makes it the campaign metric of link building: a content launch or PR push should show up as a velocity step within weeks, long before the referring-domains total or a domain-authority score visibly moves.

Weekly net figures are noisy, so the standard reading is bars for gained and lost with a rolling four-week average over the net line. Direction and acceleration are the signals; the absolute number is site-specific and only comparable to your own history.

What data does it need?

• An ahrefs_backlink_changes table with one row per gained or lost referring domain — target_domain, stat_date, referring_domain, and change_type ('gained' or 'lost') — synced daily. Normalize that value on the way in: every query below, and on the backlink growth dashboard, assumes those two labels.
• Optionally domain_rating_source per change row, so velocity can be cut by the quality of the domains flowing in and out.
• Campaign and launch dates to annotate on the chart — velocity is most useful read against the work that was supposed to move it.

SQL patterns

SELECT
  date_trunc('week', stat_date) AS week,
  COUNT(*) FILTER (WHERE change_type = 'gained') AS gained,
  COUNT(*) FILTER (WHERE change_type = 'lost') AS lost,
  COUNT(*) FILTER (WHERE change_type = 'gained')
    - COUNT(*) FILTER (WHERE change_type = 'lost')
    AS net_new_domains
FROM ahrefs_backlink_changes
WHERE target_domain = 'yourdomain.com'
  AND stat_date >= CURRENT_DATE - INTERVAL '6 months'
GROUP BY 1
ORDER BY 1;

WITH weekly AS (
  SELECT
    date_trunc('week', stat_date) AS week,
    COUNT(*) FILTER (WHERE change_type = 'gained')
      - COUNT(*) FILTER (WHERE change_type = 'lost')
      AS net_new_domains
  FROM ahrefs_backlink_changes
  WHERE target_domain = 'yourdomain.com'
    AND stat_date >= CURRENT_DATE - INTERVAL '12 months'
  GROUP BY 1
)
SELECT
  week,
  net_new_domains,
  ROUND(
    AVG(net_new_domains) OVER (
      ORDER BY week
      ROWS BETWEEN 3 PRECEDING AND CURRENT ROW
    ),
    1
  ) AS rolling_4wk_avg
FROM weekly
ORDER BY week;

Pitfalls

Where this metric applies

• Ahrefs + Metabase — the new-and-lost referring-domain feed this metric is built on
• SE Ranking + Metabase — daily new and lost backlink counts per monitored domain
• Semrush + Metabase — backlink summary snapshots to cross-check velocity against a second index

Related

Metrics

Dashboards

FAQ