RE-ENGAGEMENT Optimize Workflow Approach

Question

Hello,Can anyone please let me know if the mentioned approach for the query is correct, or if there is another way to optimize the process?Query:1. Set up a re-engagement workflow in Klaviyo for a segment of 150k(approx) unengaged contacts who haven't opened or clicked the email in the past 90-210 days.2. Throttling: Since the list size is large, batch send emails to 5K-10K contacts per day instead of sending to all contacts at once. 3. Send 6 emails with a 5-day delay between each email.Current Approach:Trigger:	Contacts enter the flow when they are added to the Unengaged (90-210 days) segment.(Note: All existing contacts in this segment should receive the email. Kindly check on the Add Past Profiles)		Batching Using Conditional Splits:	The first conditional split picks a small random percentage (4%) of contacts.		If "Yes", they receive Email #1 (Batch 1) immediately.		If "No", they wait 1 day and try again.		Repeating the Process Until All 150K Contacts Are Processed:	The same logic continues for 25 days, selecting a new batch (4%) daily.		Each batch gets its email (Batch 1, Batch 2, Batch 3... up to Batch 25).		After Receiving Email #1:	A profile property update tags each contact as part of their respective batch.		The flow ends for now, but they will continue receiving the next emails (Email #2, #3, etc.) in separate steps.		After each email, a condition will check who has clicked. Those who have clicked will exit the flow.(pending - yet to set this up in a current workflow)I have also attached screenshot for the reference. Let me know if any details are required.

Byrne C · Accepted Answer

Hi ​@DishaM13,Your flow setup looks mostly correct, but I noticed one thing you can fix. If you continue to use a 4% random sample in each subsequent conditional split, the number of people who go down the YES path will get lower each split. For example, only 3.84% of the total will go down the YES path of the second conditional split, because it’s taking 4% of the 96% who went down the original no path. This will result in an increasingly smaller number of people going down the YES path of each split. I’d recommend bumping the 4% up to 5% after a couple splits, and then up to 6%, and so on, to avoid this.After people leave your flow, are you going to create new flows to send emails 2 through 6? If so, make sure to put a filter on those flows to ensure that the people entering haven’t opened or clicked an email in the original flow.Once you take all these into consideration, your setup should work! Let me know if I can answer any additional questions.-Byrne

MANSIR2094 · Answer

Hello ​@DishaM13,Your approach is solid, but here are some optimizations for better efficiency and deliverability:Use Flow Filters Instead of Add Past Profiles – Instead of adding past profiles, set a flow filter to target unengaged contacts dynamically. This avoids sending to suppressed or inactive profiles.Optimize Batching – Instead of 4% daily over 25 days, consider using Klaviyo's "Send in Batches" feature in campaign settings for better automation and control.Engagement Exit Criteria – Ensure that profiles exit the flow if they engage at any point, preventing unnecessary emails.Gradual Warm-up – Start with smaller batches (e.g., 2.5K) and gradually increase to maintain sender reputation.Resend Logic – If a contact doesn’t open multiple emails, consider removing them earlier to avoid deliverability issues.Would you like help setting up engagement-based exits?

RE-ENGAGEMENT Optimize Workflow Approach

2 replies

Reply

Reply

Related topics

After a order is place how do I copy the customer email into a new list?

Welcome series customers Vs non customers clarification needed.icon

Which Filters are best (future proof) for Welcome Email Series?icon

Sending email as a trigger for a flowicon

integrating klaviyo with shopify to get discount emails to customersicon

Sign up

Use your Klaviyo credentials

Log in to the Community

Use your Klaviyo credentials

Scanning file for viruses.

This file cannot be downloaded