Migration10 min read

Shared to dedicated IP: the warm-up you can't skip

A dedicated IP gives you control over reputation but starts at zero. The 4-week ramp from your existing shared volume looks easy on paper. Most migrations crash at week 3 when senders push volume too fast.

Switching from a shared IP pool to a dedicated IP is one of the most counterintuitive deliverability decisions a sender can make. You move from infrastructure with established reputation (shared with hundreds of other customers, for better and worse) to a blank slate where every send during the first month builds — or breaks — your standing with Gmail, Outlook and the rest. Done well, the dedicated IP becomes your highest-leverage deliverability asset. Done badly, you spend three months recovering from the migration.

TL;DR

Dedicated IPs make sense above roughly 100,000 sends per day. Below that volume, you cannot generate enough engagement to keep the IP warm. The standard 4-week warmup ramps from 50/day to your full volume; week 3 is where most migrations fail because senders jump volume too fast. Split traffic between old shared IP and new dedicated through the warmup, monitor Postmaster daily, and accept that February-March is not the time to do this if your peak season is approaching.

When dedicated actually makes sense

A dedicated IP is worth the trade only if you can keep it warm. The threshold most ESPs cite is around 100,000 sends per day, but the more honest number is 100,000 engaged sends per day. If your list of 500,000 has 20% engaged recipients (100,000), you can run a meaningful daily cadence to the engaged subset and let the IP build healthy reputation. If you have 500,000 recipients and only 50,000 engaged, the dedicated IP starves for engagement signals between sends and reputation drifts neutral.

Below the threshold, shared IPs win because their pooled reputation is built on aggregate engagement from multiple senders. Even your weak-engagement campaigns benefit from the pool's overall health. Trying to dedicate too early means trading a stable shared baseline for a volatile blank slate.

Above the threshold, dedicated wins because your reputation is your own. A bad sender in your shared pool cannot drag you down. Postmaster Tools shows your IP reputation specifically, not a pool average. You can run consistent IP-warming and growth without external noise.

Why blank IP equals blank slate

On day one of a dedicated IP, mailbox providers see a new IP with no recipient history. Gmail's Postmaster Tools shows "insufficient data" for reputation. Yahoo and Outlook similarly default to a neutral starting position. Send too much too fast and they assume a spam burst — the same volume profile that would be fine on a warm IP triggers throttling, deferral, or outright rejection on a cold one.

Reputation builds through engagement signals over time. Replies, opens, clicks, folder moves into Primary — each one is a small positive vote. Complaints, spam-button marks, deletes-without-open are negative votes. The ratio determines reputation trajectory. In the first 4 weeks, a strict positive-skew is required because absolute volume is low and any negative event has outsized weight.

This is why the first 4 weeks are dedicated to your most engaged recipients — they reliably produce positive signals at high ratios. Weak segments must wait until week 5+, after the IP has banked enough reputation to absorb mixed engagement without sliding backward.

The 4-week ramp schedule

Standard ramp from cold dedicated IP to full volume, assuming a target peak of 200,000/day. Adjust proportionally for higher or lower volumes:

  • Week 1, day 1-3: 50/day to your top 1% engaged. Verify deliveries reach inbox; check Postmaster for early data.
  • Week 1, day 4-7: 200, 500, 1000, 2000/day. Same top engaged segment.
  • Week 2: 4k, 6k, 9k, 14k, 20k, 30k, 40k. Expand to top 10% engaged.
  • Week 3: 55k, 70k, 90k, 110k, 130k, 150k, 170k. Expand to top 30% engaged.
  • Week 4: 180k, 190k, 200k, then steady at 200k. Begin including next engagement tier.
  • Week 5+: hold steady at peak volume. Begin rotating in older segments at controlled pace.

ESPs offering dedicated IP options usually have a built-in warmup mode that enforces these limits automatically. Use it. Manual override of warmup limits is the most common cause of week-3 reputation collapse.

Splitting traffic during warmup

During warmup weeks, your full sending volume continues through the existing shared IP for everything not assigned to the dedicated IP. The split:

  • Dedicated IP: warmup-volume sends to engaged segments, following the ramp schedule above.
  • Shared IP: all remaining recipients — the bulk of your list — and any send that exceeds the dedicated IP's daily allowance.

Most ESPs allow per-campaign IP selection or per-segment IP routing. Configure routing rules so the assignment is automatic — manual per-send IP selection invariably results in mistakes that overload the dedicated IP early in warmup.

After week 5, gradually migrate remaining segments to the dedicated IP over 2-3 additional weeks, ending with the weakest engagement tier. Total migration time from start of warmup to full dedicated operation is 6-8 weeks. Anyone promising faster is selling a shortcut that damages reputation.

The week-3 cliff: where most migrations fail

Week 3 is the failure window for almost every dedicated IP migration we see. The pattern: weeks 1-2 go smoothly (low volume, no problems), the team gets confident, week 3 they push volume harder than the schedule recommends or include weaker segments earlier than planned. Postmaster reputation drops from emerging-Medium to Low. The next week's sends arrive at a worse-than-baseline state because the IP is now actively distrusted.

Recovery from week-3 collapse takes 4-6 additional weeks of restricted sending to the smallest engaged segment. Total cost is 10-12 weeks of compromised deliverability instead of the 6-8 weeks a clean migration takes. The revenue cost of this over a quarter is usually 5-10x the cost of doing the migration right the first time.

Two specific week-3 mistakes to avoid:

  • Volume jumps over 50%. Day-over-day increases should stay under 50% throughout warmup, never doubling. A jump from 70k to 150k looks normal on paper and catastrophic to mailbox providers.
  • New segment introduction. Adding a previously- excluded segment in the middle of the week resets the engagement profile. New segments wait until the start of week 4 or later.
Postmaster monitoring is non-negotiable

Daily checks of Google Postmaster Tools for the dedicated IP and sending domain throughout warmup. Set a calendar reminder. Reputation drops are visible 24-48 hours before they affect actual placement; catching the drop early lets you pull back volume and recover. Catching it late means you are already in the cliff.

Postmaster signals to watch

Three Postmaster Tools metrics matter most during warmup:

  • IP reputation: should move from "not enough data" to "Medium" by end of week 2 and "High" by end of week 4. Stuck at Low through week 3 indicates engagement is too weak for the volume.
  • Domain reputation: tracks engagement signals across all IPs sending from your domain. Should rise steadily through warmup. A divergence (domain rising, IP not) means the dedicated IP is the bottleneck — too much volume relative to engagement signal.
  • Spam rate: the user-reported complaint rate. Must stay below 0.1% throughout warmup. Above 0.3% triggers Gmail throttling regardless of other reputation indicators.

Outlook's SNDS (Smart Network Data Services) provides equivalent IP-level data for Outlook and Microsoft 365 deliverability. Sign up before warmup starts. Watch the "filter result" column for any movement away from green.

Rollback plan

If week-3 cliff or worse occurs, the rollback is to revert all sending to the shared IP, pause dedicated IP sending entirely, and let the dedicated IP "cool off" for 2-3 weeks. After cooling, restart warmup from week 1 with stricter engagement gating. Trying to push through a damaged warmup never works.

Keep the shared IP active for 8 weeks after planned cutover specifically as a rollback safety net. Most ESPs will not penalise you for keeping shared access alongside dedicated; the cost is the dedicated IP add-on, not the shared infrastructure. After 8 weeks of stable dedicated operation, the shared IP can be released.

Frequently asked questions

Can I run dedicated IP for transactional and shared for marketing?

Yes, and for many senders this is the right architecture. Transactional gets the dedicated IP to ensure password reset and order confirmations always reach the inbox. Marketing rides shared infrastructure where the pool absorbs occasional engagement variance.

What if my volume varies wildly day to day?

Dedicated IP wants steady daily volume within roughly 30% variance. If you send 500k on Tuesday and 50k on Wednesday, the IP's engagement signal looks erratic. Either smooth volume across days or stay on shared until your sending pattern is more uniform.

How does dedicated IP interact with my domain reputation?

Domain reputation is built per-domain regardless of IP. Moving to a dedicated IP does not reset domain reputation, but it does mean the IP's engagement signals start influencing domain reputation in the new combination. A weak IP with a strong domain underperforms a strong IP with the same domain.

Should I get multiple dedicated IPs?

Above 1 million sends per day, multiple dedicated IPs let you split streams (marketing, transactional, lifecycle) and isolate risk. Each IP needs its own warmup. Below 500k/day, one dedicated IP is plenty — multiple IPs at low volume each split the engagement signal and slow warmup.
Related reading

Check your deliverability across 20+ providers

Gmail, Outlook, Yahoo, Mail.ru, Yandex, GMX, ProtonMail and more. Real inbox screenshots, SPF/DKIM/DMARC, spam engine verdicts. Free, no signup.

Run Free Test →

Unlimited tests · 20+ seed mailboxes · Live results · No account required