Outlook Algorithm Changes & Cold Email in 2026: What You Need to Know

Outlook just made changes that most cold email senders missed. And if you're sending any volume to Outlook addresses - which you probably are - this matters.

Here's the situation: Outlook's filtering algorithm got more aggressive about engagement patterns. That means your campaign structure needs to adjust, or your deliverability takes a hit. This isn't about individual email quality anymore - it's about how Outlook reads your sending pattern across your entire domain and IP.

Let's break down what changed, why it matters, and what you actually need to do about it.

What Outlook Changed and Why It Matters

Outlook's algorithm now tracks sending velocity and engagement patterns more closely than before. Specifically:

The system flags accounts that send to large numbers of non-responders in rapid succession
It monitors your domain reputation across Outlook's entire network - not just individual mailboxes
It's looking at the ratio of delivered emails to actual opens and clicks, not just bounce rates

This is different from Gmail's approach. Gmail focuses on individual folder placement and sender metrics, while Outlook is essentially asking: "Does this sender's mail pattern look legitimate?"

The practical impact: You can have perfect authentication (SPF, DKIM, DMARC all passing) and still get throttled if Outlook sees your pattern as "spray and pray."

The Real Numbers You Need to Know

Here's what we're seeing in production campaigns:

Optimal send velocity: No more than 50-80 emails per day from a single warming domain to Outlook addresses when you're under 30 days old. After 30 days of consistent engagement (15%+ open rate minimum), you can push to 150-200/day.
Engagement threshold: Outlook now requires at least 12% open rate before the algorithm stops penalizing velocity. Below that, you're fighting an uphill battle with throttling.
Reply impact: A single reply resets your "sender pattern" in a positive way for roughly 48 hours. That's not anecdotal - it's measurable in bounce rate changes across warm-up periods.

What doesn't matter anymore (this changed): Having a clean list. Outlook doesn't care if your list has 5% bad emails or 15% bad emails, as long as your engagement rate stays above the threshold. They're filtering on sender behavior, not list quality.

How to Structure Your Campaign for Outlook's Algorithm

Here's the framework that works right now:

Phase 1: Domain Warm-Up (Days 1-30)

Send small, high-engagement batches only. Target 50 emails/day maximum, and segment these into groups:

Batch 1 (8am): 15 emails to your warmest, most engaged list
Batch 2 (12pm): 20 emails to secondary engagement list
Batch 3 (3pm): 15 emails to new prospects

Space out the batches by at least 2 hours. This prevents Outlook from seeing a sudden spike in your sending volume.

Phase 2: Ramp-Up (Days 31-45)

If you're hitting 12%+ open rate, increase to 120-150/day using the same spacing strategy. Add one more batch, but keep your best engagement segment separate from new cold prospects.

Phase 3: Scale (Day 45+)

Once you have 14+ days of consistent 13%+ open rates, you can hit 200-300/day. But keep doing this: never send more than 40% of your daily volume to completely cold, never-engaged addresses. Always mix in some replies, some warm replies, some engaged accounts.

Concrete example of daily distribution at full scale:

60 emails: Prior replies + engaged accounts
80 emails: Recent opens (last 7 days)
120 emails: Cold prospects from good list
40 emails: Strategic follow-ups

Subject Line Strategy for Outlook Filtering

Outlook's algorithm also looks at subject line patterns. This is new and specific. If Outlook sees similar subject lines going out in rapid succession to multiple addresses, it flags that domain.

Don't do this:

"Quick question about [Company]" (to 50 people in one batch)

Do this instead - rotate subject lines every 3-5 emails within a batch:

"Quick question about [Company]"
"[Company] + [Your Service] question"
"Quick thought about [Company]'s recent [event]"
"Following up - [Company]"
"Helping [Company] with [outcome]"

This looks natural. It looks like individual emails, not a batch.

The Unsubscribe Header Issue

Outlook now tracks unsubscribe patterns. If recipients are unsubscribing at more than 0.5% per batch, Outlook assumes you're sending unsolicited mail. That sounds low, but here's why it matters:

One unsubscribe per 200 emails is fine. One per 100 is a warning sign. When you hit one per 50, Outlook starts throttling your domain across the board - not just to that recipient, but to your entire Outlook sending.

This means your list quality matters more in 2026 than it did before, specifically for Outlook. You need to be targeting actual decision makers who have some reason to care about your message.

Monitoring and Adjustment

Check these metrics every 3 days during your warm-up period:

Outlook-specific open rate: Should be 10%+ by day 15. If it's below 8% by day 20, pause and adjust your messaging.
Inbox placement (Outlook accounts): Use a tool like 250ok or similar to check. You want 85%+ inbox placement. If you're below 75%, you're getting caught in junk or bulk folders.
Reply rate (Outlook only): This should be 0.5-1% of sends by day 25. If it's below 0.3%, your message isn't resonating with your target.

If Outlook placement drops below 80% for three consecutive days, reduce your daily volume by 50% and go back to Phase 1 spacing for a week. Outlook remembers these patterns, and recovery takes time.

What This Means for Your Overall Cold Email Strategy

The bigger picture: cold email deliverability is now multi-layered. You can't just fix one thing and assume you're good. Outlook requires you to think about sending velocity, engagement patterns, list quality, and message relevance all at the same time.

This actually makes cold email harder to just "set and forget," but it also makes it easier to stand out. Most people still send like it's 2023 - one domain, maximum velocity, minimal segmentation. If you follow the framework above, you'll have significantly better results than 80% of people sending cold email right now.

The downside: this requires active management. You can't just load a list into lemlist and walk away. You need to monitor Outlook metrics specifically, adjust velocity based on engagement, and segment your sends in a way that looks natural to the algorithm.