What the Deliverability Score Measures
Every mailbox provider — Gmail, Outlook.com, Yahoo, and the enterprise gateways behind them — chooses where to place an incoming message using the same handful of upstream signals. None of them publish the exact formula, but they all read the same inputs: did the message authenticate, how have recipients engaged with prior messages from this sender, what does the complaint history look like, is the sender on any real-time blocklist, and what is the historical pattern of the IP and the domain. The deliverability score rolls every one of those inputs into a single 0–100 number, weighted by how heavily the providers weigh them — the same composite the providers maintain, surfaced in a per-component dashboard anyone can run.
The point of having a composite is to make the decision obvious. You do not have to know whether Gmail cares more about engagement than Yahoo does to act on a dropping score — you only have to know which sub-component is dragging the line down. A high composite means the per-message decision the provider makes is a default-yes for Primary inbox; a low composite is a default-no for Junk or rejection. The score is not a measure of inbox placement on a single campaign — that is a separate inbox placement test — but it is the input the placement decision is computed from.
The Five Sub-Components
The composite is the sum of five sub-scores, each also on a 0–100 scale. The weights vary by provider, but the inputs are the same — and most scoring services use a weighting close to the one below, with authentication heaviest because it is the cheapest to keep clean.
1. Authentication (25–30%)
SPF checking against your published record, DKIM signature verification, and DMARC alignment between the authenticated domain and the visible From: header. The cheapest sub-component to keep clean: a passing check on every message adds up over weeks; a single missing record on a vendor subdomain can drop the sub-score in a day. Validate every record with the free SPF / DKIM / DMARC checker.
2. Engagement (20–25%)
Recipient-side signals: opens, replies, click-throughs, and moves-from-spam. Engagement is the only sub-component where the sender controls both halves — the content side (do recipients recognise and act on the message) and the list side (are there enough recipients on the list who will engage). Engagement is what warm-up ramps produce, and the first thing to drop on a list-hygiene regression.
3. Complaint Rate (15–20%)
Spam complaints as a share of delivered messages. Gmail treats 0.1% as the safe ceiling for bulk senders, with 0.3% the point at which throttling begins. Microsoft uses SNDS-reported complaint rate per-IP. Yahoo reports counters in its Feedback Loop. A complaint-rate spike above 0.1% over a single campaign can drag the composite down faster than any other sub-component.
4. Blocklist Status (15–20%)
Listing on Spamhaus, Barracuda, Cisco Talos, SenderScore, SORBS, or Spamcop. A single Spamhaus SBL or DBL listing can drop the composite by 10–20 points in one reporting window, because so many corporate gateways consult Spamhaus upstream of the per-message filter. Check status daily with the free blocklist checker — the cheapest signal to monitor continuously.
5. IP / Domain History (10–15%)
How long the sending domain has been active, how long the IP has been associated with the domain, and whether the domain has any past history of abuse. A brand-new domain on a fresh dedicated IP carries no negative history but has no positive history either — it reads as low-confidence and starts the warm-up curve from there. Domain age is the most lagging input; lift by minutes-per-day disciplined sending, not by configuration changes.
Run the free AI deliverability verdict on any domain
Paste a sending domain, get a 0–100 composite plus the per-sub-component breakdown and the top 3–5 risk factors driving the score. No account required. The verdict is the same composite the providers maintain on the same inputs.
Run the free verdict →How to Read the Score Over Time
A single point reading is useful for triage — paste a domain, get the composite, see what is dragging it down. The diagnostic value compounds when you read the score daily: the trend tells you whether the most recent campaign moved the composite up or down, and which sub-component reversed. Most scoring services surface a 7-day rolling composite and a 30-day rolling composite alongside the per-component breakdown, so the same dashboard answer two questions at once: is the domain healthy right now, and is the trend moving the right direction?
The dominant pattern in a dropping score usually points at the cause. A Postmaster domain reputation drop in the same window as a 0.3% complaint-rate spike usually means the most recent campaign carried the spike. A Spamhaus DBL hit on the same day as a Postmaster drop usually means the listing is the cause, not the campaign. A score that drops sub-component-by-sub-component over weeks usually means warm-up is the wrong shape or the list is going stale — neither has a single fix; both need a curve correction.
One score over time also shows what successful recovery looks like. After the underlying cause is fixed, the composite climbs back sub-component-by-sub-component, with authentication recovering fastest (hours to days after the configuration change), engagement recovering next (1–2 weeks of clean sends), blocklist status recovering as the delist clears and TTLs expire (24–72 hours), and IP / domain history recovering last (weeks of consistent clean sending).
How to Lift a Dropping Score
The fix loop is the same every time. Read the per-component breakdown to identify the lowest-scoring sub-component, run the matching free checker to confirm the underlying regression, apply the fix that recovers that sub-component, then repeat. Most senders can lift a 60 composite to above 80 in 4–6 weeks by following this pattern strictly; the slowest lift is usually the IP / domain history band, which only catches up with consistent volume and engagement.
If authentication is the smallest sub-score: run the SPF / DKIM / DMARC checker, fix the failing record (the most common cause is the vendor-subdomain misalignment — your ESP signing with d=mailvendor.net while the visible From: is your brand), and the sub-component usually recovers inside an hour of DNS propagation.
If complaint rate is the smallest sub-score: pause non-essential sends, prune non-engaged recipients, audit the most recent campaign for known spam-trigger patterns (URL shorteners, all-caps subjects, image-only HTML), and re-pace volume to a curve matched to your known-good baseline. The sub-component recovers in 1–2 weeks of clean sending if the underlying cause is caught early.
If blocklist status is the smallest sub-score: run the per-list blocklist checker or use the sender domain or IP blacklist checker for a 0–100 risk score, then identify every list with a hit, fix the underlying cause the list operator publishes (compromised account, spam-trap hit, recycled domain, open relay), submit the delist request, and allow up to 24 hours for DNS TTLs at receiving servers to expire before expecting recovered delivery.
If engagement is the smallest sub-score: re-pace volume on the warm-up schedule, prune stale recipients, and audit content for spam-trap patterns. The fastest single move is suppressing anyone who has not opened in the last 90 days — that is the segment pulling the engagement band down on every campaign.
If IP / domain history is the smallest sub-score: the path back is volume + engagement, not configuration. Stay on the warm-up curve, hold subscriber acquisition until engagement recovers, and accept that this band is the slowest to move. A 4–6 week stretch of clean sending at consistent volume is the typical recovery window.
See your deliverability today
Authentication is the first sub-component on the Deliverability Score card on your dashboard — break it down with the live verdict on any sending domain, then read the per-sub-component fixes against the same five inputs.