In Part 1, we covered business health monitoring. In Part 2, we tackled operations and inventory. Both of those are about whether the store is working.
This article is about whether the store is growing — and where the growth is actually coming from, which is often not where you think. Marketing without data is mostly an opinion supported by hope, and hope is not a marketing channel, no matter how much your social media manager believes in it. Sidekick won't tell you what to do, but it will tell you what's happening, which is the prerequisite for deciding anything.
A Quick Refresher on Sidekick
If you're new to the series: Shopify Sidekick is the AI assistant built into your Shopify admin, accessible from the chat icon in the top-right of admin (and in the mobile app). It has direct access to your store's order, customer, and product data, so its answers are based on what's actually happening rather than general advice the internet had lying around.
For marketing specifically, Sidekick has one significant limitation worth knowing up front: it sees Shopify's data well, but it doesn't have full visibility into your ad platforms (Meta, Google Ads, TikTok). The prompts below are designed to work with what Shopify does know — attributed revenue, customer behavior, channel performance as Shopify reports it — and to tell you when you need to cross-check against your ad platform's own version of events. Spoiler: the two versions will not match. They never do.
How to Use This Article
Same structure as Parts 1 and 2: each prompt comes with why it matters, how to use the answer, and when to run it. Marketing prompts in particular benefit from running on a steady cadence and writing the numbers down somewhere durable. A quarter of comparable monthly data is worth more than the single most beautiful dashboard you only look at twice.
Looking for the full reference list? See our Complete Sidekick Prompt Cheat Sheet.
The 10 Marketing & Customer Insight Prompts
1. Understand Customer Acquisition
Prompt: How many new customers did I acquire this month, and which channels did they come from?
Why it matters: Knowing how many new customers walked in the door — and which door they used — is the foundation of every other marketing decision. Without this number, every "the campaign is working" claim is technically unfalsifiable, which is convenient for whoever ran the campaign.
How to use it: Compare against last month and the same month last year. A channel doubling its share is interesting whether it's growing the total or just stealing from another channel. Pair this with ad spend by channel (from your ad platforms) to estimate a rough cost per new customer. Rough is fine — better an honest estimate than a precise attribution model nobody trusts.
When to run it: Monthly, on roughly the same day each month so the comparisons are clean and the trend is real.
2. Calculate Customer Lifetime Value
Prompt: What is the average lifetime value of my customers? How much does a typical customer spend across all their orders?
Why it matters: LTV tells you what you can afford to spend on acquisition. Most merchants don't actually know it and quietly assume it's higher than it is, which is how acquisition spend gradually bankrupts otherwise healthy stores — slowly, then all at once.
How to use it: Use LTV as your acquisition ceiling, not your floor. If LTV is $80, paying $60 to acquire a customer is acceptable only if the rest of your unit economics are healthy. If LTV is low, the leverage is in raising it (repeat rate, AOV, subscription) before scaling acquisition. Pouring money on top of a leaky funnel produces a wet funnel, not a fixed one.
When to run it: Quarterly. LTV is slow-moving; recalculating more often produces noise, not insight.
3. Measure Email Performance
Prompt: How much revenue came from email marketing this month? What percentage of total revenue was that?
Why it matters: Email remains the highest-ROI channel for most e-commerce stores, and also the most consistently underused. If email is below 20% of revenue, you have a clear, cheap, fixable opportunity sitting in front of you, and probably an ESP subscription you're not getting your money's worth out of.
How to use it: Break the number down further: how much from automated flows (welcome, abandoned cart, post-purchase) versus campaigns (newsletters, promotions)? If campaigns dominate, your flows aren't doing their job. If flows dominate but campaigns are flat, you have a list you're not talking to.
When to run it: Monthly. Also after launching any new flow, to see whether it's actually contributing or just exists.
4. Evaluate Promotion Effectiveness
Prompt: Compare revenue during my last sale event to the same period without a sale. Did the promotion actually increase total revenue?
Why it matters: Most promotions feel successful because orders spike. The honest question is whether you made more total revenue than you would have without the discount — or whether you just sold the same volume at lower margin to customers who were going to buy anyway. The honest answer is sometimes painful.
How to use it: Look at three numbers together: total revenue, total units, and average discount per order. If revenue went up but units stayed flat, you discounted customers you didn't need to discount, which is the most expensive way to feel good about a promotion. Adjust the next one accordingly — smaller discount, tighter conditions, or a different structure entirely.
When to run it: After every major promotion. Don't trust your memory of how it went; the spike is genuinely misleading, and human beings are biologically optimistic about spikes.
5. Identify Top Customer Segments
Prompt: Who are my most valuable customers? Show me customers who have placed more than 3 orders or spent over $500 total.
Why it matters: Your top customers are simultaneously your highest-margin orders, your most likely reviewers, your best source of referrals, and your most useful research panel. Treating them like a generic segment is the most expensive habit you can develop, and many do.
How to use it: Save this group as a real Shopify segment and actually use it: early access to launches, an occasional thank-you (a code, a handwritten note, a small gift), and a periodic survey. The survey is the most valuable part — these customers will tell you things you cannot get any other way, including things you'd rather not hear.
When to run it: Quarterly to refresh the segment, but the segment itself should be active continuously, not sitting in a folder being noble.
6. Analyze Geographic Performance
Prompt: Which states or countries generate the most revenue? Are there regions where I get traffic but few sales?
Why it matters: Geography surfaces problems that look invisible at the aggregate level. A region with high traffic and low conversion almost always has a specific, identifiable barrier — shipping cost, delivery time, tax surprise, or simply a competitor who happens to be down the road from those customers.
How to use it: For underperforming regions, check shipping rates from the customer's perspective (use an incognito browser and a real ZIP — yours doesn't count). If rates are the problem, test a region-specific shipping offer for one month. For overperforming regions, consider geo-targeted ads to lean into demand you've already proven exists.
When to run it: Quarterly. Geography doesn't shift quickly, and monthly checks mostly produce noise that looks like signal.
7. Find Cross-Sell Opportunities
Prompt: Which products are most often purchased together in the same order?
Why it matters: Affinity data is some of the most useful information in your store, and somehow consistently among the least used. The pairs your customers naturally buy together are your most credible bundles, cross-sells, and "frequently bought together" recommendations — credible because they're based on actual behavior, not your guesses about behavior.
How to use it: Pick the top three pairings and act on each differently: build one into a bundle with a small discount, add one as a cross-sell on the product page, and add one to a post-purchase email. Measure each independently after 30 days so you know what worked, rather than waving generally in the direction of "bundling."
When to run it: Quarterly. Pairings are surprisingly stable, and acting on them takes longer than discovering them anyway.
8. Monitor Review Sentiment
Prompt: What is my average product rating? Which products have the lowest ratings, and what are customers saying about them?
Why it matters: Reviews are the most visible quality signal on a product page, and a single low-rated product can drag down the perception of the whole store. They're also unusually honest — customers will tell strangers things they'd never tell you directly, especially if you've ever responded defensively to feedback in an email.
How to use it: Read the actual review text on the bottom three products. Complaints cluster: "runs small," "photos don't match," "smaller than expected," "arrived damaged." Each has a different fix (page copy, photos, packaging, supplier). Pick the cluster, fix it, and check back in 30 days. Don't fix all three at once or you won't know which one worked.
When to run it: Monthly. New negative reviews are most damaging when they're recent and visible, which is most of the time they exist.
9. Compare New vs. Returning Behavior
Prompt: What is the average order value for first-time customers vs. returning customers?
Why it matters: In a healthy store, returning customers spend more per order than first-timers. They trust you, they've removed the size or fit uncertainty, and they're more likely to add the second item to the cart. If your returning AOV is the same as or lower than your first-time AOV, your retention strategy is not really a retention strategy — it's a routine.
How to use it: If the gap is small, look at whether your post-purchase flow is doing any actual selling, or just being polite. A discount-heavy retention strategy (constant 15% codes) often produces lower returning-customer AOV than before retention efforts began — a bad outcome wearing the costume of engagement.
When to run it: Quarterly. Pair it with the retention rate from Part 1, prompt #5 for the fuller picture.
10. Measure Campaign ROI
Prompt: How much revenue came from paid advertising this month, and what was my return on ad spend (ROAS)?
Why it matters: Ad spend without ROAS tracking is just a steady outflow of cash with extra steps. The number doesn't have to be excellent — it has to be honest, and it has to clear your gross margin plus overhead. Anything else is technically just enthusiasm.
How to use it: Calculate your real breakeven ROAS once and write it down somewhere you can find it later (roughly 1 ÷ your gross margin). Every campaign gets measured against that, not against the platform's preferred metric. Cross-check Sidekick's number against the ad platform's own attribution — they will disagree, and the truth is usually somewhere between them, leaning toward whichever one you trust less.
When to run it: Monthly at minimum. Weekly if you're actively scaling — but don't make decisions on less than two weeks of data, no matter how confident the dashboard seems.
Marketing & Insights Routine
Marketing data benefits from comparison across time, which means the cadence matters as much as the prompt. Pick a schedule and stick to it; the value is in the trendline, not the snapshot.
Weekly:
- #3 Email revenue
- #10 Ad ROAS
Monthly:
- #1 New customers by channel
- #4 Promotion effectiveness (after each promo)
- #8 Review sentiment
Quarterly:
- #2 Customer lifetime value
- #5 Top customer segments
- #6 Geographic performance
- #7 Cross-sell patterns
- #9 New vs. returning behavior
Wrapping Up the Series
Across these three articles we've covered thirty prompts:
- Part 1: Business health monitoring — sales, conversions, retention
- Part 2: Operations & inventory — stock, fulfillment, margins
- Part 3: Marketing & customers — acquisition, campaigns, insights
Sidekick is a tool, not a strategy. Its value depends entirely on what you ask it, how often, and whether you actually do something with the answer. Start with the prompts most relevant to your current pain, build a routine you'll actually keep (not the ideal one you'd build for someone else), and let the data do the arguing in your weekly review.
All thirty prompts in one place: Sidekick Prompt Cheat Sheet.
CTA: Need help turning insights into strategy? Let's talk about growing your Shopify store.