AI personalization and retention: transforming SMS and on-site experience into growth drivers

Executive summary
Download the short version of this case study, with the key metrics, client outcomes, and roadmap. 

[⬇ Download executive summary PDF]

At a glance

• Industry: Specialty apparel (direct-to-consumer fashion)
• Region: United States (Austin, TX HQ, selling nationwide)
• Size: ~90 employees, $28M annual revenue
• Commerce stack: Shopify Plus, Amazon Seller Central (secondary channel)
• Marketing stack: Klaviyo, Attentive (SMS), Nosto (personalization)
• Key challenges: flat repeat purchases, poor SMS ROI, generic on-site experience, no churn prediction capability

The challenge: repeat revenue stuck in neutral

Despite strong growth in acquisition through Meta Ads and TikTok campaigns, this U.S. apparel brand saw customer lifetime value (CLV) plateau. Leadership depended on continuous ad spend to fuel topline growth, but profitability was suffering.

• Flat repeat purchases: the repeat customer rate held steady at ~28%, even after heavy investments in promotions.
• SMS fatigue: Attentive campaigns felt impersonal. Customers complained about “spammy” texts, and CTR was below 2%.
• Generic on-site experience: personalization limited to static “bestsellers,” leading to disengagement.
• Churn blindness: executives had no way to identify at-risk customers. Retention campaigns were reactive and often triggered after customers were already lost.

As the VP of eCommerce put it:

“We’re great at getting customers in the door. But we’re not good at keeping them. We can’t afford to rely on ads forever — we need personalization and intelligence.”

Our solution: personalization beyond campaigns

Every engagement started with structured discovery by the Growth Architecture Center (GAC) — bringing together commerce strategists, CRM experts, data analysts, and compliance leads.

The GAC led interviews with the VP of eCommerce, CRM manager, and IT lead. Three critical priorities emerged: personalize SMS, make the on-site experience relevant, and lay the foundation for AI-driven churn prediction.

Delivered solution

Phase 1, 12 weeks:

• SMS personalization: rebuilt Attentive workflows using Shopify + Klaviyo data. Instead of batch messages, campaigns were segmented by browsing history, loyalty tier, and purchase anniversaries.
• On-site experience: implemented Nosto’s advanced personalization features to show bundles (“frequently bought together”) and dynamic homepage carousels tailored to each visitor.
• Unified data foundation: integrated Shopify Plus, Klaviyo, and GA4 into Snowflake, building customer-level profiles.
• Dashboards: deployed Looker Studio for executive-ready views of repeat purchase rates, CAC vs CLV, and SMS ROI.

Planned / Future (Phase 2–3):

• Churn prediction models: developing machine learning models in AWS SageMaker to score customers by churn likelihood.
• Win-back automation: linking churn scores back to Klaviyo/Attentive for targeted reactivation campaigns.
• Cross-channel orchestration: extending personalization logic into paid ads (TikTok and Meta custom audiences).

Business impact

The impact of this project went beyond numbers — it reshaped how the client approached retention and customer relationships.

• SMS performance improved significantly: CTR rose from ~2% to ~6%, while unsubscribe rates declined by about 25% as messages became more relevant.
• Repeat purchases began to rise: within three months, the returning customer rate grew from 28% to just over 31% — a modest but material six-figure annual impact at current scale.
• On-site engagement improved: average time on site increased by ~20%, and AOV grew by ~8% in segments where personalized bundles were displayed.
• Executive visibility unlocked: leadership gained dashboards showing revenue split between new and returning customers. For the first time, the CFO could align marketing spend directly with retention-driven revenue.

Not everything was perfect:

• Forecasting churn with AI is still a work in progress.
• Nosto personalization sometimes “misfires” (e.g., recommending out-of-season items), requiring human oversight.
• Training the CRM team to manage data flows in Snowflake took longer than expected, requiring several weeks of workshops.

But the client valued that we were open about these realities. 

As the VP of eCommerce said:

“This isn’t magic — it’s process, data, and discipline. Now we see a path to growth that isn’t entirely dependent on ad spend.”

Why this partnership worked

• Quick wins first: SMS personalization delivered visible ROI within eight weeks, proving value early.
• Scalable foundation: Snowflake data warehouse and AWS SageMaker roadmap mean the client can grow without replatforming.
• Teams were enabled to run Nosto/Attentive workflows themselves, with GetDevDone acting only as an escalation partner when needed.
• Transparency: challenges (like personalization misfires) were addressed openly with mitigation plans, building trust.

Future growth roadmap

• Phase 2: deploy churn prediction models in SageMaker and link scores to Klaviyo/Attentive win-back campaigns.
• Phase 3: activate personalized audiences in TikTok and Meta ads to align acquisition with retention data.
• Phase 4: explore digital product passports (anticipated U.S./EU compliance for fashion by 2026).

This roadmap ensures that the client’s eCommerce strategy doesn’t just chase short-term sales, but evolves into a predictable, AI-driven retention engine.