Dedicated IP vs. Shared IP for Email: What the Data Actually Shows

The dedicated IP question comes up in almost every deliverability conversation I have with ecommerce email teams. It’s one of those decisions that feels high-stakes — and it is — but it’s also one of the most misunderstood levers in email deliverability. The question isn’t just “dedicated or shared?” It’s “dedicated or shared, at what volume, with what domain reputation, and are you prepared to manage it?”

InboxEagle analyzed 15,174 IPs — 7,166 dedicated and 8,008 shared — across ecommerce sending programs. The results are more nuanced than the conventional wisdom suggests, and the data tells you exactly where the real differences live.

What the Numbers Are Actually Telling You

You’d expect a dedicated IP to outperform a shared IP across the board. Lower spam, higher inbox. The data tells a more interesting story — and the explanation is in the numbers themselves.

Dedicated IP senders in the dataset do run a cleaner spam rate: 11.99% versus 13.22% for shared IP senders. That 1.23 percentage point gap is real. But look at where those emails are going instead of spam: dedicated IPs route 68.77% to the Promotions tab, versus 65.69% for shared IPs. That 3+ point Promotions gap directly accounts for the lower Primary inbox rate (19.24% dedicated vs. 21.09% shared). The emails aren’t going to spam — they’re going to Promotions.

And that Promotions gap has consequences the spam-rate number hides entirely: mail in Promotions is delivered, uncomplained-about, and frequently unread. Inbox vs. promotions vs. spam covers why optimising against spam rate alone can quietly cost you revenue.

The Volume Gap That Explains the Whole Result

One number in this dataset does more explanatory work than any other, and it is easy to skim past.

Dedicated IPs averaged 16.33 emails per IP. Shared IPs averaged 46.43 — nearly three times as much.

That is not a finding about IP types. It is a finding about who is on them. A dedicated IP with 16 emails against it is not a mature sending program; it is a sender who recently made the switch and is somewhere in the middle of a warm-up. Low volume, short history, few accumulated engagement signals. Shared IPs, meanwhile, carry ESP infrastructure that has been building reputation for years across thousands of senders.

So the counterintuitive headline resolves cleanly once you read the cohorts rather than the categories. Dedicated IP senders in this data are cleaner but younger. Shared IP senders are borrowing from a larger reservoir of established trust. The 3-point Promotions gap is what a reputation still being built looks like from the outside.

Which means the honest reading of this study is not “dedicated IPs are worse.” It is that the comparison is measuring a warm-up cohort against an established one, and you should expect your own numbers to look like the dedicated column for the first several weeks after any move — because that is what the dedicated column mostly is.

The infrastructure decision that usually matters more, and costs nothing: which subdomain you send from. Root-domain senders in our data run a 26.0% spam rate against 13.3% for subdomain senders — a far larger gap than anything in this table, available without changing IPs at all.

The Case for Staying on a Shared IP

If you’re sending fewer than 100,000 emails per month, a well-managed shared IP pool is almost certainly better for your deliverability than starting fresh on a dedicated IP.

ISPs treat a new IP like a bank treats a borrower with no credit history — cautious, low limit, watch-and-see. Your ESP’s shared IPs don’t have that problem. They’ve been warming for years, sending millions of emails, and building relationship data with every major inbox provider.

The risk on shared IPs is real but often overstated. Yes, if another sender on your pool has a complaint spike, it can affect your deliverability. But good ESPs segment senders into pools by quality tier. SendGrid’s shared IP pool architecture automatically routes higher-performing senders into cleaner pools — you’re not randomly pooled with spammers, you’re pooled with senders who perform at a similar level to you.

The more important takeaway: if your deliverability is struggling on a shared IP, the problem is almost never the IP. It’s your domain reputation, your list quality, or your engagement rates. Switching to dedicated doesn’t fix any of those.

The Case for a Dedicated IP for Email

The argument for a dedicated IP is about control and isolation, not a guaranteed performance lift. Your sending history is yours alone. A bad month from another sender doesn’t touch you. Your complaint rates, engagement patterns, and sending consistency write your own reputation story with no outside interference.

For high-volume ecommerce brands — large retailers, brands running heavy seasonal campaigns, agencies managing multiple high-volume accounts — a dedicated IP gives you a reputation asset you fully own. You can monitor it, optimize it, and protect it without any exposure to anyone else’s sending behavior.

The volume thresholds from major ESPs give you a useful calibration. Klaviyo requires 1,000,000+ marketing emails per month before qualifying a sender for its dedicated IP program. Postmark recommends 300,000+. The consensus across providers sits around 100,000 consistent monthly sends as the minimum where a dedicated IP starts to make practical sense — and shared vs. dedicated IP has the platform-agnostic version of this framework if you are weighing it across ESPs.

Note what those thresholds imply. If your ESP will not sell you a dedicated IP, that is not an obstacle to route around — it is the provider telling you your volume cannot sustain the reputation a dedicated IP would need.

The operative word is consistent. An ecommerce brand sending 400,000 emails in November for BFCM and 30,000 in January is a worse dedicated IP candidate than one sending 100,000 every single month. ISPs watch for dormant IPs — long gaps in sending erode the reputation you’ve built. Consistency of volume matters more than peak volume.

IP Warming: The Part Most Brands Skip

Move to a dedicated IP and you cannot start at full volume. Warming means ramping gradually over 4–8 weeks so ISPs register positive engagement before you hit them with the full list — and skipping it means a snap judgement about an IP they have never seen, which usually resolves to spam folder, throttling or blocking. Recovery from a botched warm-up is 4–8 weeks of careful low-volume sending to your best subscribers, which is what you should have done to begin with.

The week-by-week ramp is in IP and domain warming; the warm-up planner generates one for your volume, and Klaviyo senders should follow the Klaviyo warm-up process for the platform specifics. If your reason for moving is a seasonal volume spike rather than a permanent increase, eCommerce email reputation covers handling the spike without changing IPs at all.

Shared vs. Dedicated IP: The Decision Framework

One nuance worth flagging: if your deliverability issues trace back to list quality or authentication — which is the cause in the majority of cases — neither IP type will fix it. Solve those first, then evaluate IP type.

Domain Reputation Drives More Than IP Type at Gmail

Here’s the context that reframes the whole dedicated vs. shared debate for the modern sending environment: Gmail has decisively shifted toward domain reputation as its primary filtering signal, ahead of IP reputation. Mailgun’s analysis confirms it consistently — strong domain reputation with average IP reputation beats perfect IP reputation with weak domain reputation.

Microsoft Outlook still weights IP reputation heavily through SmartScreen, which is why IP type matters more for your Outlook audience. Yahoo considers both roughly equally. IP reputation vs. domain reputation breaks down which provider leans which way and why the balance has been shifting.

Microsoft’s 2025 change tilted this further toward domain. Bulk mail to Outlook, Hotmail and Live.com now has to pass SPF, DKIM and aligned DMARC or it is refused outright with a 550 5.7.515 error — an authentication gate that sits in front of reputation entirely. No IP is clean enough to survive failing it, and no dedicated IP will buy you past it. See Outlook bulk sender requirements.

What this means in practice: your email authentication setup and your engagement signals have far more leverage over placement than IP type does. As covered in the state of email authentication in eCommerce, 94% of brands have the records but far fewer enforce them. A dedicated IP on a domain with weak authentication will underperform a shared IP with clean auth and an engaged list. Every time.

Get authentication right first. Then your sending domain structure. IP strategy comes third — and for most brands, never.

The Bottom Line

The InboxEagle data across 15,174 IPs makes it clear: dedicated IPs aren’t inherently better, and shared IPs aren’t a compromise. They serve different senders at different stages.

• Dedicated IPs have lower spam rates (11.99% vs. 13.22%) — cleaner senders, but lower volume and newer sending history
• Shared IPs have slightly higher Primary inbox (21.09% vs. 19.24%) — because dedicated IPs route 3+ more points to Promotions during reputation-building phases
• The avg_emails_per_ip gap (16.33 vs. 46.43) is the key variable — dedicated IP senders are warming; shared IP senders are established
• Below 100K/month: shared IPs are almost always the right call — borrowed ESP reputation beats a fresh IP with no history
• Above 100K/month, sending consistently: dedicated IPs give you full control — but only after 4–8 weeks of warming done properly
• Domain reputation and authentication outweigh IP type at Gmail — solve those first, then revisit IP strategy

Whatever you’re on today, monitor your actual inbox placement across all major ISPs — not just delivery rate. Our Q1 2026 Gmail benchmark data shows 21.6% of ecommerce emails still land in spam regardless of IP type, and the deliverability checklist is the place to start if you haven’t audited your full setup recently.

See how your sending program actually performs across ISPs →

Sources

• InboxEagle native research data — 15,174 IPs (7,166 dedicated, 8,008 shared), Q1 2026 inbox placement analysis
• Klaviyo Help Center — Dedicated vs. Shared IPs
• Mailgun — Domain Reputation vs. IP Reputation: What Gmail Cares About
• SendGrid — Shared IP Pools 101
• Postmark — Dedicated IP Addresses
• Google Email Sender Guidelines

Note: Content created with the help of AI and human-edited and fact-checked to avoid AI hallucinations.