Einstein STO and Scoring: The Two Einstein Wins for Most Accounts

Einstein has six features; two produce consistent measurable lift on almost every account:

• Send Time Optimization - 5-15% open rate improvement from per-subscriber send timing
• Engagement Scoring - better segmentation with Journey Builder Scoring Splits

Here's how to deploy both.

Send Time Optimization setup

Concept: instead of sending to 50k subscribers at 9 AM sharp, send each subscriber at their individual optimal time - when they historically open most.

Implementation in Journey Builder:

One Activity added. That's the entire setup.

The STO Activity evaluates each entering subscriber's open history and holds them in the journey until the time of day they're most likely to open. Some get sent at 7 AM; some at noon; some at 10 PM.

STO: when it works and when it doesn't

Works well:

• Large audiences with varied time zones
• Subscribers with some engagement history (STO needs data to learn from)
• Non-time-sensitive content (newsletters, promotions)

Doesn't work well:

• Time-sensitive content ("Sale ends at midnight" - can't shift delivery to the next morning)
• New subscribers without history (STO has nothing to learn from)
• Small audiences where individual timing doesn't aggregate into meaningful signal

For time-sensitive sends, don't use STO. Pick a send window and accept uniform delivery.

STO optimization window

STO can be configured to send within a window - e.g., "optimal time between 8 AM and 6 PM." Useful when you don't want emails going out at 3 AM even if the subscriber historically opens then.

Default window often works; tune if the business requires it.

Engagement Scoring setup

Concept: SFMC calculates a 0-100 engagement score per subscriber based on historical opens and clicks with your brand. In Journey Builder, branch based on score.

Implementation in Journey Builder with Scoring Split Activity:

Three paths, same journey. Each subscriber takes one based on their score.

Patterns with Engagement Scoring

Pattern 1: Frequency filter

High-volume promotional send. Use Scoring Split to drop low-engagement subscribers from this send - they're unlikely to engage and risk unsubscribing from frequency fatigue.

Low-score subscribers aren't removed from the account - just from this particular send. They still get reduced-frequency content.

Pattern 2: Tiered content

Welcome series uses Engagement Scoring on email 4+:

• High-engagement branch: deep product content, upsell offers
• Mid-engagement: standard educational content
• Low-engagement: re-engagement survey, "Are you still interested?"

Pattern 3: Re-engagement trigger

Daily automation checks for subscribers whose score drops below 30. They enter a re-engagement journey automatically.

Data STO and Scoring need

Both features require engagement history. New accounts with no send history don't have data for Einstein to learn from.

Minimum guidelines:

• 60-90 days of consistent sending
• At least 10,000 subscribers with at least one open or click
• Ideally 3-6 months of history before trusting STO/Scoring in production

For new SFMC deployments, don't activate these in the first month. Let history build, then turn them on.

Mistake 1: STO on time-sensitive sends

"Sale ends at midnight" sent with STO. Some subscribers get it at 1 AM - after the sale ended. Disaster.

Rule: no STO on time-sensitive content.

Mistake 2: Scoring Split without path definitions

Team adds Scoring Split but only defines the High path. Medium and Low subscribers fall into a dead end. Silent failure.

Always define all three paths explicitly.

Mistake 3: Trusting Einstein output stale

Subscriber's habits change. Einstein model lags. Review STO-recommended times monthly; if the audience behavior has shifted, give the model fresh data or recalibrate.

Measuring impact

Before/after comparisons to quantify the lift:

STO test:

• Control: standard 9 AM send to 50% of audience
• Test: STO Activity enabled for the other 50%
• Metric: open rate difference
• Typical result: 5-15% lift on STO path

Engagement Scoring test:

• Control: standard cadence to full audience
• Test: Scoring Split cuts low-engagement subscribers from high-frequency promo sends
• Metric: unsubscribe rate per send + total sends to healthy subscribers
• Typical result: lower unsubscribe, stable revenue (the low-engagement subscribers weren't converting anyway)

Rollout plan

Month 1-2: observe existing performance, establish baseline. Month 3: deploy STO on one recurring non-time-sensitive campaign. Measure. Month 4-5: deploy Scoring Split on one welcome or re-engagement journey. Measure. Month 6+: expand to other campaigns and journeys where patterns apply.

Don't bulk-convert. Test incrementally.

Takeaway

STO and Engagement Scoring are the two Einstein features that produce consistent measurable wins. STO is one Activity added to a Journey. Scoring is one Split Activity plus three paths. Both require a few months of history to be effective. Deploy incrementally, measure lift, expand to other campaigns. 5-15% open rate lift from STO alone pays for Einstein's licensing on most accounts.

---

Deploying Einstein STO and Scoring on client SFMC accounts? Our Salesforce team designs and measures Einstein feature rollouts with before/after testing on production engagements. Get in touch ->

See our full platform services for the stack we cover.