Shared IP or Dedicated IP for Klaviyo: When Each Makes Sense

Most Klaviyo deliverability decisions do not matter as much as people think. The choice between a shared IP and a dedicated IP is one that genuinely does.

For smaller senders, the shared IP pool that Klaviyo assigns by default is almost always the right answer. For larger senders with consistent volume and a clean list, a dedicated IP isolates your reputation from your neighbours. The hard part is knowing which side of that line you sit on, and the answer depends on more than just list size.

This guide walks through the actual decision criteria, the volume thresholds that matter, and what happens when senders move in either direction at the wrong time.

What shared IPs actually do

When you send through Klaviyo on a shared IP, your messages leave from an IP address that is also sending for dozens or hundreds of other Klaviyo customers. Mailbox providers evaluate that IP's reputation as a single unit. If the other senders on that IP are sending clean, engaged, well-authenticated mail, the IP carries a strong reputation and your messages benefit from it. If a few neighbours run aggressive re-engagement campaigns or buy lists, the IP's reputation drops and your delivery suffers alongside theirs.

Klaviyo does active pool management. Their deliverability team segments senders into pools based on behaviour and reputation indicators, so a well-behaved account is not randomly placed next to a list buyer. This is one of the genuine strengths of their platform and is often underappreciated. But pool management is not the same as isolation. A bad actor in your pool still drags the pool's reputation, and you still sit in that pool until Klaviyo rebalances it.

For most ecommerce senders sending fewer than 100,000 messages per month, this tradeoff is worth it. The pool average is usually stronger than what a single brand could build on its own, and you do not have the warming burden that comes with a dedicated IP.

What dedicated IPs actually do

A dedicated IP is an IP address that sends mail only for your brand. Mailbox providers evaluate that IP's reputation based solely on your sending patterns. If you send consistent volume, maintain a clean list, and generate positive engagement, you build a reputation that reflects your own behaviour and nothing else.

The benefit is isolation. If another sender on a shared pool does something reputation-damaging, you are unaffected. If your own reputation drops, it drops because of your sending, which means the diagnostic path is clear and the fixes are within your control.

The cost is responsibility. A dedicated IP starts with no reputation. Mailbox providers have no history on it, and until you build that history through consistent, engaged sending, your delivery will be worse than it would be on a well-warmed shared pool. Warming takes weeks. During that time, you are sending meaningful volume into a reputation vacuum, and recovery from a warming mistake takes longer than the warming itself.

A dedicated IP also requires sustained volume. If you go quiet for a week, reputation decays. Mailbox providers expect a regular signal from an IP, and gaps in sending are interpreted as either an unreliable sender or an IP that has been reallocated.

The volume thresholds that actually matter

The rough industry guideline is that a sender needs at least 100,000 to 200,000 messages per month to a single mailbox provider before dedicated IP reputation becomes measurable. Below that threshold, there is not enough volume for Gmail or Microsoft to form a stable reputation picture. You are effectively invisible to their reputation systems, which means you neither benefit from a good reputation nor suffer from a bad one. You just sit at the platform's default treatment.

In practical terms:

• Under 50,000 sends per month: dedicated IPs almost always hurt. You do not have the volume to build reputation, and you lose the lift that a healthy shared pool provides.
• Between 50,000 and 250,000 sends per month: it depends. If your sending is consistent week over week, your list is clean, and your engagement is strong, a dedicated IP can work. If your volume is spiky, if you rely on promotional campaigns that concentrate on specific days, or if your engagement is marginal, a shared IP is safer.
• Above 250,000 sends per month: dedicated usually makes sense. You have enough volume to form reputation at every major ISP, and the isolation benefit outweighs the warming cost.

These are not hard cutoffs. They are calibration points. The actual decision needs to account for how concentrated your volume is at specific providers. A sender with 300,000 monthly sends that are 80% to Gmail and 15% to a few other major providers has enough volume for dedicated reputation at Gmail and not much elsewhere. That sender sits in a fine place for shared pools with excellent Gmail performance, or a dedicated IP if Gmail isolation is the specific goal.

When dedicated solves nothing

Senders frequently move to a dedicated IP hoping it will solve a deliverability problem. In most cases, it will not. The most common cause of Klaviyo deliverability issues is not IP reputation. It is authentication misconfiguration, list hygiene, or sending patterns that trigger filtering regardless of which IP is used.

A common pattern: a sender moves to a dedicated IP after a month of Gmail inbox placement issues, the dedicated IP does not help, and the underlying cause turns out to be a DKIM alignment issue that was already there. Moving to a dedicated IP just transfers the authentication problem to a new IP, where it produces the same filtering behaviour. The fix is a DNS correction, not an IP change.

Before moving to dedicated, verify:

• Your DKIM and SPF are aligned and signing correctly at every sending domain you use in Klaviyo.
• Your DMARC policy is published and your reports are clean.
• Your list is receiving normal bounce rates (under 2% hard bounce on engaged segments, under 5% on broad sends).
• Your complaint rates are below 0.1% at Gmail and 0.3% at Microsoft.
• Your sunset policy is actively suppressing non-engaged subscribers.

If any of these are broken, a dedicated IP will not help. Fix them first. If they are all healthy and you still see deliverability issues, and you have the volume to support dedicated, then it is worth considering.

When shared stops working

The other direction matters too. There are cases where a sender has outgrown shared pools and is being held back by neighbours. The signals for this look different from authentication or list hygiene issues:

Your domain reputation at Gmail Postmaster Tools (available once you have a branded sending domain) looks clean, seed tests show consistent inbox placement at the IP-owner level, and bounce codes are healthy, but your inbox placement still drifts. Note that as a Klaviyo customer you cannot access SNDS or IP-level GPT data for Klaviyo's IPs, so the IP-side picture has to come from Klaviyo's deliverability team or third-party monitoring.

You see consistent, small delivery fluctuations that do not correlate with anything you changed.

Your peers at similar volume and engagement levels, using different tools or dedicated infrastructure, see meaningfully better placement at the same ISPs.

In these cases, a dedicated IP can genuinely help, because the drag on your delivery is coming from neighbour behaviour that you cannot influence. This is rarer than senders assume, but it does happen, especially for brands with very specific engagement profiles that do not match the rest of a shared pool.

The warming question

Any move to a dedicated IP requires warming, which is a separate topic we cover in its own guide. The short version: you cannot just move your full volume onto a new dedicated IP on day one and expect normal delivery. Mailbox providers will treat the IP as suspicious, filtering will increase, and recovery from a failed warming can take six to twelve weeks of sustained, disciplined sending.

Warming schedules that work involve ramping volume per ISP over four to eight weeks, prioritising the most engaged segments of your list first, and monitoring delivery at each provider individually. If you do not have the operational capacity to watch this carefully, the move to dedicated is likely to make your deliverability worse before it makes it better.

The decision framework

The useful question is not "shared or dedicated." It is "what am I actually trying to solve, and will this change it."

If your goal is reputation isolation from a specific problem you can trace to pool neighbours, and you have the volume and consistency to warm a new IP, dedicated is a reasonable move.

If your goal is to fix a deliverability problem with an unclear root cause, stop. Diagnose the actual problem first. A dedicated IP that solves nothing adds warming risk to whatever you were already dealing with.

If your volume is under 100,000 monthly messages, stay on shared. The platform's pool management is better than anything you can build at that scale.

If your volume is above 250,000 monthly and concentrated at two or three major providers, dedicated is probably worth the investment, assuming your underlying hygiene is sound.

Between those, it is a judgment call, and the specific shape of your sending matters more than a single threshold.

Summary

A dedicated IP is a tool for senders with enough volume, consistent sending patterns, and clean underlying infrastructure to build and maintain their own reputation. It is not a fix for deliverability problems caused by authentication, list, or content issues. Shared pools work well for most senders under a certain volume, and Klaviyo's pool management is genuine value that smaller brands often underestimate.

Make the decision based on what you are actually trying to solve, not on a vague sense that dedicated is more "serious." Dedicated is more exposed, not more protected, unless the preconditions are in place.