Write the growth equation in business terms
Revenue alone can hide expensive growth. Start with qualified sessions, conversion rate, average order value, gross margin, fulfillment and return costs, customer acquisition cost, repeat purchase, and contribution over a useful time horizon.
Break the numbers down by new versus returning customer, product or category, channel, geography, promotion status, and acquisition cohort. A channel that looks weak on first-order ROAS may acquire durable customers, while a high-ROAS campaign may depend on discounts or existing demand.
- New and returning customer revenue are separated
- Discount, return, fulfillment, and margin effects are visible
- AOV and units per order are reviewed by product and offer
- Repeat purchase is measured by cohort rather than a blended total
Trace the promise from channel to product experience
Review whether search results, ads, creators, email, and social content set expectations the store can fulfill. Then test navigation, collection pages, search, product comparison, product detail, variants, sizing, shipping, returns, cart, and checkout support on mobile.
The goal is not to add more persuasion everywhere. It is to give the customer the right information at the moment they need it, preserve the brand, and remove uncertainty that does not help anyone.
- Landing context matches the campaign or search promise
- Products are easy to find by the way customers think
- Product pages explain use, fit, benefits, evidence, delivery, and returns
- Offers and bundles are understandable without hidden conditions
- Mobile performance and interaction do not obstruct purchase
Audit what happens after the first order
Retention begins with product and delivery experience. Review expectation setting, order communication, packaging, onboarding or usage guidance, support, review requests, replenishment timing, cross-sell logic, loyalty, and winback.
Lifecycle messages should respond to customer context rather than simply fill a campaign calendar. Use product, order, engagement, and timing data to make the next message useful.
- Welcome and abandonment flows reflect customer intent
- Post-purchase communication helps customers receive value
- Replenishment timing matches plausible product use
- Cross-sell recommendations make sense with the original purchase
- Winback is measured against customer value and discount cost
Verify the data behind every conclusion
Test product views, list views, add-to-cart, checkout, purchase, refund, promotion, search, and key lifecycle events. Check currency, value, item identifiers, duplicate purchase events, consent behavior, referral exclusions, UTMs, and platform discrepancies.
Build a decision view that combines store truth, analytics behavior, ad spend, lifecycle performance, product and cohort data, and margin where available. Attribution remains imperfect, so use multiple evidence sources and state the limits.
Choose one constraint and a balanced experiment set
Rank opportunities by likely contribution impact, confidence, effort, customer value, and risk. Balance quick repairs with deeper learning: fix broken tracking and mobile friction, but also test product story, merchandising, creative, offers, and lifecycle assumptions.
A good 90-day roadmap is small enough to ship, diverse enough to learn, and connected to the economics the team actually needs to improve.