Mogo / Senior Marketing Operations Manager / 2024 – Present
Winback and cross-sell at scale
900K dormant users across two products and one list. I treated it as a multi-iteration system, not a campaign.
The read
There was one combined list of about 900K people: users who had gone inactive on Intelligent Investing, and lending clients who had borrowed through Mogo but never touched the investing side at all.
Two completely different products sat inside that one list. Manage is automated weekly investing into an S&P 500 portfolio. Self-Directed is manual stock and ETF trading. A single generic reactivation email would have ignored that difference and undersold both.
The lending segment wasn't a reactivation problem at all. They'd never invested through Mogo, so this was a cross-sell, not a winback, even running through the same list and the same lifecycle system.
Nobody had treated this list as more than one audience for one campaign. That was the actual gap.
The decision
Run it as a multi-iteration system, not a single send, with each iteration doing double duty as a campaign and a list hygiene pass. List quality was unknown, so pruning hard upfront risked cutting people who'd have converted later. The first pass had to surface real signal before I touched the list.
I split messaging by product fit instead of one offer to everyone: a contribution calculator, modeling weekly amounts out to age 75, for Manage-fit users. Fiscal.ai, the research tool already included in every membership, for Self-Directed-fit users. Lending clients got a third frame: you've used Mogo to borrow, here's how to start building.
The metric I chose
Conversion into meaningful Intelligent Investing activity, tracked cumulatively across iterations. Not opens, not clicks. I also used Iteration 1's number as a decision gate, since it told me how to treat the list going into Iteration 2: who to keep, who to cull, who to move onto a low-touch drip instead of suppressing entirely.
The build
Built the segmentation model in Braze using product fit, lapse recency, and prior engagement. Iteration 1 ran the full 900K from February to June 2025: re-introduction, product-specific hooks, a one-click objection survey for non-converters, then a light taper for non-responders.
After Iteration 1, I culled hard bounces, spam complaints, and repeat soft bounces, bringing the reachable list to about 600K. For the segment that never engaged at all, I built a low-frequency monthly drip instead of going dark, since a sudden re-engagement blast later would read as a cold list to inbox providers. Iteration 2 ran July 2025 through January 2026 against that cleaned 600K.
~2.7%
Iteration 1 conversion, 900K audience
~5%
Iteration 2 conversion, cleaned 600K audience
~54K
Total conversions across both iterations
The read held. Segmentation outperformed volume: conversion nearly doubled between iterations without sending more, because the list was cleaner and each hook matched the right product. The cross-sell worked because it acknowledged the existing lending relationship instead of ignoring it. Iteration 3 is now running against the remaining ~300K, with a sunset program planned for whoever's left after that.
Tech
Braze, segmentation and Canvas orchestration
Microsoft Dynamics, CRM source and lending history
Snowflake, cohort and behavioral data
n8n, multi-step automation and segment sync
Next: The behavioral trigger layer
Miguel N. Monzones
Vancouver, BC, Canada
