How to improve email deliverability: a nine-step plan that holds up

Delivery and deliverability are two different numbers, and mixing them up is how senders convince themselves nothing is wrong. Delivery — the rate your platform reports — counts messages a receiving server accepted. A message filed straight into the junk folder counts as delivered. Deliverability is the number that pays: the share of your mail that reaches an inbox a human actually looks at.

Your platform can't see that second number, which is why "99% delivered" dashboards coexist with collapsing open rates. Improving it is not one trick. It's a sequence, and the sequence matters: authentication before reputation work, list repair before engagement tactics, monitoring before you trust any of it. Here is the order we'd run.

If you haven't diagnosed why placement is poor, do that first — the causes and their evidence are covered in a separate guide. This article assumes you're ready to fix.

Step 1: authenticate — SPF, DKIM, DMARC, all three

Nothing else on this list works while authentication is broken, because mailbox providers now treat it as table stakes. Google's bulk sender guidelines and Yahoo's sender requirements both demand SPF, DKIM and a published DMARC policy from anyone sending at volume, and both filter unauthenticated mail harder at every scale.

Concretely: publish one SPF record covering every service that sends as your domain, enable DKIM signing with your own domain in the d= field (not your platform's), and publish a DMARC record — p=none with a reporting address is a legitimate start. Then verify, don't assume: the email authentication checker runs all the checks against your live DNS in one pass. The full record-by-record walkthrough lives in our guide to the three protocols.

Step 2: send from a subdomain you can afford to learn on

Send marketing from mail.yourdomain.com or news.yourdomain.com, not the bare root your invoices and password resets depend on. Reputation accrues per domain; separating streams means a rough patch in marketing doesn't drag transactional mail down with it. If you're already sending from the root domain with a long history, don't switch just for the sake of it — a new subdomain starts with no history at all and needs warming.

Step 3: fix the list before you touch anything else

Reputation is a weighted average of how recipients react to your mail, so the fastest way to raise it is to stop sending to the people dragging it down. Three moves, in order of impact:

1. Remove hard bounces immediately and permanently. Repeatedly hitting dead addresses is a classic marker of a careless sender.
2. Suppress non-engagers. Anyone with no opens or clicks in roughly six months belongs in a re-permission campaign or off the list. Shrinking your list this way routinely raises revenue per send — you lose only the audience that had already left.
3. Never import purchased or scraped addresses. One bad import can contaminate a clean list with spam traps and dead accounts, and the damage lands on your domain's record, not the list broker's.

Going forward, confirmed (double) opt-in keeps the problem from rebuilding. It costs some list growth and buys you a list where every address demonstrably wanted in.

Step 4: get your complaint rate under control

User-reported spam is the strongest negative signal a mailbox provider has. Google's Postmaster Tools documentation states the threshold plainly: keep user-reported spam below 0.3%, and treat 0.1% as the working target. The levers are unglamorous — set expectations at signup, send at the frequency you promised, make unsubscribing effortless (one-click list-unsubscribe is a requirement under the bulk sender rules, not a courtesy), and stop mailing segments that never engage. Anyone who wants out of your list should find leaving easier than reporting you.

Step 5: warm up anything new

New domain, new IP, new platform — receivers have no model of you, and unfamiliar senders at volume look like exactly one thing. Ramp volume gradually, starting with your most engaged recipients so early sends generate opens and replies rather than silence. The full schedule logic is its own guide, but the principle fits in a sentence: never send a volume tomorrow that your history today can't justify.

Step 6: keep volume and cadence boringly consistent

Filters model your normal and penalize deviation. A sender who mails 50,000 twice a week is more predictable — and better placed — than one who alternates silence with 400,000-message bursts. If a spike is coming (product launch, seasonal peak), build toward it over a few sends rather than jumping, and hold cadence steady rather than doubling frequency in the week that matters most to you.

Step 7: make the content match the promise

Content weighs less than reputation in modern filtering, but it still weighs. Keep a sensible text-to-image balance (never image-only), make link text match link destinations, skip URL shorteners, and write subject lines the message actually delivers on. Curiosity-gap subjects harvest opens and pay for them in unsubscribes and spam reports — the metrics that outlast the campaign. A plain-text-looking email from a trusted domain outperforms a beautiful one from a distrusted domain every time.

Step 8: monitor, because everything above decays

Deliverability work doesn't stay fixed; records break, lists age, reputation drifts. The minimum monitoring kit:

• Google Postmaster Tools — domain reputation, spam rate and authentication trends for your Gmail traffic. Free, and it only collects data from the day you register, so register early.
• DMARC aggregate reports — the receiver-side view of your authentication health across all providers. Reading them is its own skill.
• Placement spot-checks — mailboxes you control at Gmail, Outlook and Yahoo, checked with real sends, not assumptions.
• Configuration re-checks after every change — a two-minute pass through the email deliverability score whenever DNS or platform settings move.

Step 9: reassess the platform itself

Two senders with identical practices can land differently depending on whose infrastructure carries the mail — shared pool quality, throttling behavior and authentication defaults vary between providers. That's measurable, and measuring it is this site's core work: we send identical campaigns through each platform to a controlled seed list and publish the inbox rate, per provider, with the test date. If you've done steps one through eight and placement still lags, the platform belongs on your suspect list. Our test protocol explains exactly how we isolate the platform's contribution.

The order is the point

Skip-around execution is why deliverability projects stall. Content polish (step 7) on top of a broken SPF record (step 1) accomplishes nothing; a warm-up (step 5) executed onto a stale list (step 3) warms up a reputation problem. Run the sequence, verify each step with the tools rather than assumptions, and give the slow steps — reputation and engagement recovery — weeks rather than days before judging them.

Related guides

• Email authentication checker — verify SPF, DKIM and DMARC in one pass
• SPF checker — record syntax and the 10-lookup limit
• Email deliverability score — a 0–100 setup score with specific fixes
• What is DMARC? — the policy layer explained
• Domain reputation — the score you're actually improving
• Dedicated IP — when owning your reputation outright makes sense
• Why are my emails going to spam? — the diagnosis-side companion

About this guide

Written by InboxRatio Editorial. The recommendations reflect what we observe in our controlled placement testing across email platforms and in current Google and Yahoo sender documentation — not vendor marketing claims, which we treat as claims. No affiliate relationship influences the guidance here.

Methodology

InboxRatio's deliverability data comes from sending real campaigns through each tested platform to a controlled, provider-spanning seed list and recording per-message placement. This guide's ordering reflects which factors that testing and provider documentation show to matter most. The protocol is public: how we test.

Last updated

21 August 2026. Provider requirements referenced here are re-checked quarterly against Google and Yahoo sender documentation.

Frequently asked questions

How long does it take to improve email deliverability? Configuration fixes (authentication, DNS) can move placement within days. Reputation-driven recovery is measured in weeks of consistent, well-received sending. Be suspicious of anyone promising a specific date — mailbox providers don't publish recovery timelines.

What's a good email deliverability rate? There is no universal published benchmark, and vendor "99%" figures count server acceptance, not inbox placement. The useful practice is tracking your own inbox placement per provider over time with seed mailboxes, and treating any sustained decline as actionable.

Does buying a dedicated IP improve deliverability? Only for senders with enough steady volume to keep its reputation alive — and it transfers all reputation risk to you. Low-volume or irregular senders are usually better placed in a well-policed shared pool. See dedicated IP for the trade-offs.

Should I delete inactive subscribers? Doesn't a bigger list help? Delete or re-permission them. List size is not a deliverability input; engagement rate is. Every non-engager in a send dilutes the positive signals that drive placement, so a smaller, active list outperforms a large, cold one on both placement and, usually, absolute revenue.

Can I improve deliverability without changing email platforms? Usually, yes — authentication, list quality, complaint control and cadence are sender-side factors and account for most placement problems. The platform becomes the prime suspect only after those are demonstrably clean.

Do open rates still matter now that opens are unreliable? Apple's Mail Privacy Protection and image proxies inflate raw open counts, so treat opens as a trend indicator, not a truth. Clicks, replies and spam-report rates are the sturdier engagement signals for deliverability purposes.

If you take one thing from this guide: run the nine steps in order, verify each with evidence instead of assumption, and re-check on a schedule because every fix decays. Your first move is two minutes long — run the email deliverability score and see which step your domain fails today.