· 11 min read
IP warm-up for Braze — the practitioner's playbook
Most IP warm-up guides tell you how fast to ramp. Few tell you which users to send to each day — and that's where most real warm-ups come off the rails. This is the playbook I've watched work across consumer-scale CRM programs over a decade: the ramp curve, the Random Bucket Number methodology, and the discipline to not blow it up on day 10.
Justin Williames
Founder, Orbit · 10+ years in lifecycle marketing
Why a new IP has to be warmed
Mailbox providers — Gmail, Yahoo, Microsoft, Apple iCloud — track the reputation of every IP address that sends them mail. Reputation is built from consistent volume, low bounce rates, low complaint rates, and engaged recipients. A brand-new IP has none of this signal, so their filters default to suspicion: messages land in spam, get rate-limited, or bounce outright.Source · GoogleEmail sender guidelinesOfficial Gmail sender requirements covering reputation, authentication, bounce rates, and complaint thresholds.support.google.com/mail/answer/81126
A warm-up schedule sends a small volume on day one, slightly more on day two, and so on — typically exponentially — until you reach your target daily volume after two to four weeks. The goal is to let the major ISPs observe clean, engaged sending patterns long before you push real volume.
The standard ramp — and why it's exponential
A 14-day warm-up typically starts at around 2% of target volume, doubles roughly every two days, and reaches 100% by day 14. A 30-day warm-up doubles less aggressively but ends in the same place. The curve is exponential because mailbox reputation compounds — small early successes earn you the headroom for bigger volume later.
The shape, visualised:
Standard 14-day IP warm-up ramp Sends on day 1 start at ~2% of target daily volume and ramp exponentially to 100% by day 14. The curve is steeper in the back half because ISP filters reward consistent engagement signals before granting headroom for larger volumes.
The ramp target is a ceiling, not a floor. If your program peaks at 100,000 sends on Tuesdays and only 30,000 on Wednesdays, warm up to your peak. Never exceed what the ramp schedule has earned — ISPs notice volume spikes more than almost any other signal, and a single day at double the expected volume can set the warm-up back by 5–7 days.
Random Bucket Numbers matter more than ramp speed
The question "how fast can I ramp" gets most of the attention. But the more important question is "which users do I send to each day?". If day one is 2,000 highly-engaged users and day 14 is 100,000 users including the dormant and the disengaged, your reputation curve will collapse on day 14 — because you suddenly look like a different sender.
The fix is Random Bucket Numbers. Assign every subscriber a stable random integer from 0–9,999 at signup. Bucket 0–199 is a random 2% of your base, bucket 0–999 is a random 10%, and so on.Source · BrazeRandom Bucket Number attributeOfficial Braze docs on random_bucket_number — the attribute used for consistent sampling, IP warm-up cohorts, and Global Holdout Groups.www.braze.com/docs/user_guide/engagement_tools/segments/random_bucket_number
On day one, send to users in bucket 0–199. On day two, expand to 0–299. Each bucket expansion adds NEW users but does not remove old ones — so engagement patterns stay consistent across the ramp. The resulting send population is an expanding random sample: every day looks like a scaled-up version of the previous day, which is exactly what ISPs reward.
The Orbit IP Warm-Up Planner generates the exact RBN ranges for each day of the schedule and includes optional Global Holdout Group carve-outs so your measurement programs stay intact during the ramp.
What actually derails a warm-up
< 2%
Bounce rate ceiling during warm-up. Above triggers rate-limiting.
< 0.1%
Spam complaint rate. Above actively damages reputation.
> 90%
Inbox placement (seed tests). Below means reputation is fading.
Three things derail warm-ups in practice: un-engaged sending, volume spikes, and a surging complaint rate. The sequence usually looks like this. A marketing stakeholder wants to "just test one quick campaign at full volume" on day 10. Volume doubles. Bounce rate spikes because dormant users are swept in. Complaint rate follows because those dormant users don't remember subscribing. The IP gets temporarily blocked by Microsoft. The recovery is expensive.
Defensive discipline: suppress users who have been inactive for 90+ days during the warm-up window. Quarantine any list segment that came from a bulk import without explicit confirmation. Enforce the ramp schedule in writing and treat the volume ceiling as a hard rule, not a suggestion.
Monitor actively during the ramp: bounce rate should stay under 2%, spam complaint rate under 0.1%, and inbox placement (measurable via seed tests) above 90%. If any of those miss, slow the ramp or tighten the audience before proceeding. The full deliverability guide covers what to watch once the warm-up completes, and the Orbit Braze Deliverability Health Check skill pulls these numbers directly from the Braze API so monitoring is automated rather than a manual daily task.
When to warm a new IP vs use a shared pool
A dedicated IP is only worth the warm-up overhead if your sustained daily volume justifies it. Below roughly 100,000 emails per day, you're almost always better off on a shared IP pool that benefits from thousands of other senders' established reputation.
Above that threshold — and especially for regulated or high-stakes sending (transactional, financial services, high-volume promotional) — a dedicated IP gives you isolation from other senders' reputation problems. The tradeoff: you own the reputation entirely, and a week of bad sending undoes months of trust.
If your Braze plan includes multiple sending IPs, a third-way option: warm the new IP alongside an existing established one, splitting daily volume between them and gradually shifting more to the new one as reputation builds. This is the most conservative option and takes longer (typically 4–6 weeks to fully transition), but it eliminates the cliff on day 14 where the new IP has to stand alone.
The measurement that tells you it worked
The warm-up is complete when five conditions hold for at least a week of sending at target volume: bounce rate stays under 2%, spam complaint rate under 0.1%, inbox placement above 90% across Gmail and Microsoft, open rate on the warm-up cohort matches your normal sending cohort, and sender reputation scores (from Google Postmaster Tools, SNDS, and your deliverability monitor if you have one) are consistently High or Good.Source · GooglePostmaster ToolsGoogle's free tool for monitoring Gmail sender reputation, spam rate, authentication results, and delivery errors.postmaster.google.com
Skip any one of those and you risk sending at target volume with an IP that isn't actually ready. The most common version of this failure: open rate matches but inbox placement seed tests are showing 78% instead of 92%, which means a meaningful slice of your sends are landing in the Promotions tab or spam folder and not being captured in your standard opens count.
Frequently asked questions
- How long does an IP warm-up take?
- Standard warm-ups run 14 to 30 days. Fourteen days is aggressive and works if your engagement rates are strong and your list is clean. Thirty days is the conservative choice for unknown-quality lists, high daily volumes (over 500K/day), or senders who've had deliverability problems before.
- What is a Random Bucket Number in Braze?
- A Random Bucket Number (RBN) is a persistent integer from 0 to 9,999 assigned to every Braze profile. It stays constant for the lifetime of the user, which means that 'users with RBN less than 200' is always the same random 2% of your audience. This makes it the right tool for consistent sampling during IP warm-up — day one's audience stays in day 14's audience.
- Should I warm up a new IP or stay on a shared pool?
- Below ~100,000 sends per day, Braze's shared IP pool typically delivers better than a half-warmed dedicated IP because it benefits from thousands of other senders' established reputation. Above that threshold — or for regulated sending — a dedicated IP gives you isolation and justifies the warm-up cost.
- What happens if I break the ramp and send too much on one day?
- Expect a bounce-rate spike, temporary rate-limiting from Gmail and Microsoft, and damaged sender reputation for weeks. The recovery is to drop back to the last compliant day's volume for 3–5 days before resuming. In severe cases (bounce rate over 5% or complaint rate over 0.3%) you may need to restart the warm-up entirely.
- Do I need a Global Holdout Group during warm-up?
- Only if you're already running a Global Holdout Group as a measurement strategy. Holdouts don't affect warm-up mechanically — they just shrink the audience the warm-up applies to. The Orbit IP Warm-Up Planner includes an optional holdout carve-out so the RBN math stays correct when a holdout is active.
- Can I warm one IP while still sending from another?
- Yes, and it's the most conservative approach. Split send volume between the warming IP and an established one, gradually shifting more to the new one as its reputation builds. Braze supports this via IP pool configuration. Plan for 4–6 weeks total rather than 14 days — but you eliminate the day-14 cliff where the new IP has to stand alone.
- How do I know the warm-up worked?
- Five conditions should hold for at least a week at target volume: bounce rate under 2%, complaint rate under 0.1%, inbox placement above 90% across Gmail and Microsoft, open rate on warm-up cohort matches your baseline cohort, and sender reputation scores in Google Postmaster Tools and Microsoft SNDS are consistently High or Good.
This guide is backed by an Orbit skill
Related guides
Email deliverability — the practitioner's guide
Deliverability is the cumulative result of every send decision over the lifetime of a domain. This guide covers the four pillars — authentication, reputation, engagement, and list hygiene — and how to recover when one breaks.
Apple Mail Privacy Protection, four years in
In 2021, Apple broke the email open rate. Four years later, the dust has settled — and the lifecycle programs that adapted are outperforming the ones still measuring like it's 2020. Here's what actually changed.
The unsubscribe page is the most important page in your lifecycle program
The page every lifecycle team ignores is the one that quietly decides sender reputation, suppressed-list quality, and the fate of your next quarter's deliverability. A short defence of why it's worth the ten-minute rebuild.