How to Audit Your Shopify Store in 15 Minutes

Written by Gentian Shero

Co-founder & CSO at Shero Commerce

How to Audit Your Shopify Store in 15 Minutes

Welcome to the March edition of Commerce Jam.

Most Shopify stores convert at 1.4%. The top 20% convert at 3.2%. That gap rarely comes down to better products or bigger budgets. It comes down to whether someone has actually looked under the hood recently.

This edition is built around a simple premise: a 15-minute structured audit will surface more revenue opportunities than most merchants find in a quarter of paid media testing. We walk through the exact framework we use at Shero across client audits, the five categories where stores consistently leak money, and the benchmarks that tell you where you actually stand.

Beyond the audit, February was a busy month. Shopify posted record results, ChatGPT walked back in-chat checkout, and Claude Cowork.

It is all in here.

In this issue:

How to Audit Your Shopify Store in 15 Minutes
Shopify Platform Updates
AI, SEO & Agentic Commerce
eCommerce News & Industry Trends
AI Agent Testing Playbook
Shero News
Final Thoughts

How to Audit Your Shopify Store in 15 Minutes

The average Shopify store converts at 1.4%. The top 20% convert at 3.2%. That gap is not explained by better products or bigger ad budgets. It is explained by operational discipline.

Most brands do not have a structured process for catching the problems that accumulate quietly in every fast-growing store. Slow pages stay slow. Broken search stays broken. Trust signals that should be on the product page never get added. A full agency audit is expensive and infrequent. Day-to-day operations leave no room for a structured review. So the friction compounds.

Article content — Shopify auditing framework

This article gives you a 15-minute diagnostic built on the five categories where Shopify stores lose the most revenue. At Shero, we use this framework across client audits, and the same issues appear regardless of store size or GMV. Technical performance first, because it affects everything else. Backend operations last because those problems are invisible until they become expensive.

One number worth sitting with: a store that loads in one second converts at 2.5 times the rate of a store that loads in five seconds. On average, every additional second costs 4.42% in conversion rate.

Continue reading the full article on the Shero Commerce blog

Shopify Platform Updates

These updates were pulled from the Shopify Changelog and our internal team.

1. Shopify Q4 2025 earnings: B2B GMV up 96%, orders from AI search up 15x

Shopify reported Q4 2025 revenue of $3.67 billion, a 31% year-over-year increase and the highest quarterly revenue in company history. Full year 2025 revenue hit $11.6 billion, up 30%. GMV reached $123.8 billion for the quarter. The company announced a $2 billion share repurchase program effective February 17, 2026. The standout numbers for merchants: B2B GMV grew 96% in 2025 and 84% in Q4 specifically. Orders from AI searches are up 15x since January 2025, still from a small base but growing fast. President Harley Finkelstein called 2025 "Shopify at full throttle." For merchants who have been waiting for Shopify's B2B infrastructure to mature, these numbers are your signal. (Shopify Q4 2025 Results)

2. Shopify Scripts: June 30, 2026, deadline is 16 weeks away

Shopify Scripts stop executing on June 30, 2026. No grace period. Any custom checkout logic running on Scripts, including discounts, shipping rules, and payment conditions, breaks at midnight. The deadline was already extended once from August 2025. Do not count on another extension. Start with the Scripts Customizations Report in your Shopify admin to see exactly what you are running. Custom Function development runs $2,000 to $10,000, depending on complexity. Scope it now, not in May. (Shopify Help Center)

3. Legacy Customer Accounts officially deprecated: February 26, 2026

Legacy customer accounts were formally deprecated on February 26. Existing stores retain access for now but are no longer receiving feature updates or technical support. A full sunset is coming. The new system has fewer customization options out of the box. Merchants with workflows built around legacy accounts, including custom login pages, account-specific pricing displays, and loyalty integrations, need to audit those dependencies. We have active client conversations on this. It is not a simple swap for stores where loyalty programs and account-gated features depend on the old infrastructure. (Shopify Changelog)

4. Shopify POS version 11.0: faster cart and checkout experience

POS 11.0 shipped February 18 with a redesigned cart and checkout flow. Shopify describes it as faster, clearer, and more consistent. For retail teams processing high transaction volumes, checkout UI friction adds up in real ways, particularly during busy periods. Update your POS app and verify your staff is trained on the new layout before peak periods. (Shopify Changelog)

5. Shopify POS Hub now shipping: wired connections replace Bluetooth

The Shopify POS Hub went generally available on March 2 and is now shipping to pre-orders. It connects iPad and Android tablets to checkout hardware, including card readers, printers, scanners, and cash drawers, over wired USB instead of Bluetooth. For omnichannel merchants with busy retail floors, Bluetooth dropped connections during checkout are a real operational cost. Wired connections eliminate that variable entirely. If your retail setup has been Bluetooth-dependent, this is worth evaluating. (Shopify Changelog)

6. Metafields as dimensions and filters in Analytics: February 12

You can now use custom metafields, including product, variant, customer, and order metafields, as dimensions and filters inside Shopify Analytics. Segment by material, loyalty tier, product category, or any custom attribute you have built into your catalog. This eliminates a persistent workaround where merchants had to export data to external tools just to slice it by custom fields. For stores with rich product taxonomies, this is a meaningful analytics upgrade. (Shopify Changelog)

7. Sidekick can now create customers and companies: February 10

Sidekick can now create customer and company records directly from a text prompt. Describe what you need, and Sidekick fills the form. Small in isolation, but it continues the pattern of Sidekick moving from answering questions to executing admin tasks. For B2B merchants managing company accounts, reducing the manual overhead of record creation matters at scale. (Shopify Changelog)

8. Adaptive Pricing for Managed Markets: March 3

Shopify now automatically builds duties, taxes, currency conversion, and merchant-of-record fees into international prices in Managed Markets. Shoppers see a single localized price with no surprise charges at checkout. Checkout cost surprises are a primary driver of international cart abandonment. If you are running Managed Markets, verify your pricing rules are correctly configured for key markets now, before your next peak period. (Shopify Changelog)

9. Duty calculation breakdowns now visible on orders: February 26

Merchants can now see exactly how duties are calculated on individual cross-border orders and verify those calculations against their own expectations. For merchants selling internationally with complex tariff exposure, this transparency was missing. Particularly useful right now given the current tariff environment and the volume of duty-related disputes merchants are dealing with. (Shopify Changelog)

10. POS Editor: edit smart grid tiles without rebuilding

The POS Editor now lets staff edit smart grid tiles directly rather than deleting and recreating them. Lock screen and customer display customization are also now accessible from clearer layout sections. Minor update but for retail operations that frequently adjust checkout grid layouts, it removes a time-consuming friction. (Shopify Changelog)

11. Marketing automations moving on March 24: no action needed

On March 24, Shopify Messaging automations move to the Messaging app. All other marketing automations move to the Flow app. No merchant action is required and existing automations keep running. Worth flagging to your team so nobody is confused when automations appear to have relocated in the admin. Same functionality, different navigation path. (Shopify Changelog)

12. Shipping in Quick Sale: March 2

Merchants can now add shipping to quick sale carts and share a payment link. The buyer enters their address at checkout. Useful for DM selling, custom orders, and pre-launch scenarios where you need to take payment and ship without a full storefront live. (Shopify Changelog)

13. VAT number validation at checkout: February 25

EU and UK merchants using Shopify Tax can now enable VAT number validation at checkout. When a valid VAT number is entered, the reverse charge exemption is applied automatically. For B2B merchants selling to EU businesses, this removes a manual step that was creating friction and occasional errors in cross-border B2B transactions. (Shopify Changelog)

14. Hydrogen tracking cookie migration: April 30 deadline

Merchants running Hydrogen storefronts need to migrate legacy tracking cookies by April 30, 2026. Missing this deadline affects analytics attribution accuracy. Shopify says most Hydrogen stores only need a version upgrade. If you are on Hydrogen and have not done this yet, it needs to go on your list this week. (Shopify Developer Changelog)

AI, SEO & Agentic Commerce

You do not have time to track every AI development. Here are the latest developments for last month.

1. Every platform wants to own the AI shopping layer. None of them do yet.

February and March sharpened a contradiction that will define merchant strategy for the next two years. Perplexity launched shopping in the US with PayPal-integrated checkout. Shopify is powering multiple AI discovery surfaces through Agentic Storefronts. AI platforms are projected to reach $20.9 billion in retail ecommerce spend in 2026, roughly 4x the 2025 figure, at about 1.5% of total retail ecommerce. At the same time, Amazon has been expanding its robots.txt to block OpenAI, Meta, Google, and other AI crawlers from its catalog, while building its own AI shopping tools like Rufus and Buy For Me. Shopify is selectively participating, powering Perplexity's catalog while maintaining control of its checkout layer. (eMarketer)

Every platform wants to own the AI shopping layer, not be disintermediated by it. The merchants who will win are the ones keeping product data clean and structured enough to be surfaced by any discovery layer, while keeping checkout, customer data, and post-purchase relationships on their own infrastructure. Handing the entire transaction to an AI intermediary early means no leverage later. (Digital Commerce 360)

2. Reddit's AI shopping carousels: community-validated product discovery enters commerce

On February 19, Reddit announced it is testing AI-powered shopping carousels inside search results for a small group of US users. The feature pulls products from real Reddit discussions and matches them with Dynamic Product Ads partner catalogs, surfacing pricing, images, and direct purchase links for queries like "best noise-canceling headphones." Reddit's search weekly active users grew 30% year over year, from 60 million to 80 million. Reddit Answers, its AI-powered conversational search, grew from 1 million to 15 million weekly active users in 2025. (TechCrunch)

Reddit communities are among the highest-intent destinations for product research on the internet. Subreddits in tech, gaming, outdoor, home goods, and beauty categories drive purchase decisions at scale. Reddit is not surfacing random products. It is surfacing products that actual users recommend. That means authentic presence in relevant communities, not brand advertising, is the brand-building play before this channel matures. Merchants who are genuinely active in their product category subreddits now have a first-mover position when Reddit scales this feature.

3. TikTok closes US deal, tightens logistics for Shop sellers

TikTok finalized its US joint venture with Oracle, Silver Lake, and MGX in January 2026, with the three taking 50% ownership of the new TikTok US entity. With the platform's US future resolved, TikTok moved immediately to tighten its commerce infrastructure. Starting on February 25, 2026, all US TikTok Shop sellers must fulfill orders only through TikTok Shop Logistics Services. Seller Shipping, in which merchants used their own carrier accounts, is no longer available to new sellers and will be phased out for existing sellers by March 31. (TikTok Shop)

For merchants selling on TikTok Shop, this is an operational fork. Centralized fulfillment means faster delivery expectations are enforced automatically. Seller performance metrics are calculated weekly. Merchants using disconnected fulfillment tools will struggle. For merchants not yet on TikTok Shop: the platform's 200 million US users and 6% commission rate remain compelling, but the operational requirements are now materially higher.

4. ChatGPT is retreating from agentic commerce. Here is what that actually means.

On March 5, the Information broke the story: OpenAI is walking back Instant Checkout. Buyers clicking products in ChatGPT will now be redirected to the retailer's app or site rather than completing the purchase in chat. The real numbers tell you why. Of Shopify's millions of merchants, only about a dozen went live with ChatGPT native checkout. OpenAI had not even built a system to collect and remit state sales taxes as of February 2026. A clear signal that transaction volume never materialized. Shopify president Harley Finkelstein confirmed the bottleneck is on the AI side, not the merchant side.

This is the same pattern we saw with Facebook Shops and Google's Buy with Google. Both tried to own the transaction layer. Both redirected back to the merchant's site when it got hard. The fundamental tension is that AI is probabilistic and checkout is deterministic. There is no room for a confident guess when the price needs to be the price. Catalog normalization at scale is an unsolved problem. Only Google Shopping has cracked it, and it took them years.

Agentic commerce is not dead. The bottleneck is infrastructure, not intent. Shopify's Agentic Storefronts and the Universal Commerce Protocol are still the most credible bets on solving this properly. But if you are expecting AI-driven checkout to move revenue this year, recalibrate. Discoverability is the real opportunity in 2026. Transactions are coming. They are just not here yet.

5. AI traffic is under 1% for most brands. Are you even measuring it?

AI referral traffic is still under 1% of total traffic for most eCommerce brands. The Previsible State of AI Discovery Report, which analyzed nearly 2 million LLM sessions across 12 months, found AI represents just 0.13% of total sessions, roughly 1 in 769 visits. A separate academic study by Kaiser and Schulze covering 973 eCommerce sites and $20 billion in combined revenue found that ChatGPT referral traffic was approximately 0.2% of total sessions, about 200 times smaller than Google organic traffic.

The hype would have you believe AI has already taken over discovery. The data says otherwise. But the behavior pattern is worth paying attention to. AI visitors show stronger engagement. They browse, compare, and read. They have often already made their decision before they click.

That is a measurement problem, not a channel problem. Most merchants look at last-click GA data, see poor conversion from ChatGPT, and write off the channel. What they are missing is that AI is top-of-funnel. A buyer who discovers you through ChatGPT may convert three sessions later through direct. You will never connect those dots in a standard attribution model.

Build a dedicated AI traffic channel in Google Analytics now. According to Littledata, traffic from AI tools lands in the Referral channel by default because GA4 treats them as external websites, which means it gets mixed in with regular referrals, and you lose the signal entirely. Set up a custom channel group to capture sources such as ChatGPT, Perplexity, Claude, and Gemini separately. Track engagement separately. Compare AI-exposed users against all users in terms of conversion over 30 days. Treat the channel as top-of-funnel until your own data tells you otherwise. The merchants building this measurement infrastructure now will have a real advantage as traffic numbers grow.

6. Google's AI Mode and February core update: what changed for eCommerce

Google completed its February 2026 Discover core update after 21 days. AI Overviews now appear in roughly 34% of searches depending on query type. Google AI Mode, still in limited rollout, now shows contextual overlay pop-up link cards on desktop, making products more visible within AI answers. A separate analysis found that the presence of an AI Overview correlates with a 58% lower average clickthrough rate for the top-ranking page. The pattern for ecommerce merchants: transactional queries still drive clicks. Informational queries are increasingly being answered on the results page. The merchants most at risk are those who built their entire discovery funnel on informational content with weak transactional intent. (Search Engine Roundtable)

eCommerce News & Industry Trends

These are the news items that are actually worth your attention this month.

1. Stripe in acquisition talks with PayPal

Stripe, valued at $159 billion following a secondary share sale in February 2026, is in reported acquisition talks with PayPal. PayPal CEO Alex Chriss was fired in February 2026. PayPal stock is down 19% year to date. A combined entity would control approximately 65% of global online payments. (CNBC)

Shopify Payments is built on Stripe infrastructure. PayPal is one of the most widely used alternative payment methods in Shopify checkout. If this closes, the two largest payment layers in Shopify's ecosystem would be the same company. Nothing to act on today. Worth monitoring closely. The integration complexity alone would take years, but the pricing and fee implications for merchants running both are worth thinking through now.

2. Forrester: US eCommerce to reach $1.8 trillion by 2030

Forrester released a new five-year US retail forecast on February 24, 2026, projecting that eCommerce sales will reach $1.8 trillion by 2030, up from an estimated $1.5 trillion in 2025. That would put eCommerce at 29% of total US retail sales, with the remaining 71% still moving through physical stores. Forrester attributes growth to Gen Z entering the workforce with more purchasing power and improvements in fulfillment and logistics. (Forrester via Retail Dive)

The number that deserves more attention: a separate McKinsey analysis cited in the same research cycle estimates AI agents could mediate $3 to $5 trillion in global consumer commerce by 2030. That is not a contradiction to Forrester's channel data. It is a layer on top of it. The transaction still happens. The discovery and decision journey looks completely different.

3. Shopify Q4 2025: 31% revenue growth, $2B buyback, B2B the real story

Shopify posted Q4 2025 revenue of $3.67 billion, exceeding analyst estimates by 2%. Full year 2025 revenue hit $11.6 billion with a 30% year-over-year increase. The company announced a $2 billion share repurchase program effective February 17, 2026. For merchants, the signal is in the platform investment priorities: Catalog, Sidekick, UCP, and B2B infrastructure. The 96% B2B GMV growth in 2025 is where Shopify is putting real development resources. Merchants who have been waiting for the platform to mature for complex B2B are out of excuses. (Shopify Q4 2025 Results)

4. 34% of shoppers plan to spend less in 2026, 43% citing tariffs

An eMarketer consumer survey found 34% of shoppers intend to spend less in 2026, with 43% of those citing tariffs as the primary reason. Walmart reported general merchandise inflation above 3% in Q4 2025, attributed to tariff-related costs. Columbia Sportswear and Levi Strauss both announced high single-digit price increases on seasonal merchandise. (eMarketer)

For mid-market merchants, this is a conversion rate and AOV environment, not a traffic environment. The merchants who come out ahead in cautious spending periods are the ones who tighten operations and protect margins. Discounting is the wrong instinct here. Loyalty, repeat purchase frequency, and trust signal optimization are the right ones.

5. Over 1,000 tariff lawsuits filed as trade policy uncertainty continues

As of early March 2026, more than 1,000 lawsuits have been filed challenging the current US tariff policy. The legal environment around tariffs remains in flux. For merchants who built pricing and margin models around a specific tariff scenario, the uncertainty itself is the risk. Inventory decisions made in Q1 based on expected tariff levels may need to be revisited as court rulings move. (Bloomberg)

6. Amazon narrows how it enforces on time delivery rate

Starting February 28, Amazon updated the enforcement of its 90% On Time Delivery Rate threshold for seller-fulfilled orders. The threshold has not changed. What changed is how Amazon applies it. Previously, dropping below 90% put your entire seller-fulfilled catalog at risk. Now Amazon targets only the specific listings, dragging your rate down. For multi-SKU sellers, this is a meaningful improvement.

The protection mechanics remain the same: enable shipping settings automation, enable automated handling time, and use Amazon Buy Shipping labels for OTDR-protected shipments. Consistently missing 90% still puts your full catalog on the line. But for sellers close to compliance, this gives you a path to clean up problem listings without a full shutdown. If you sell across Amazon and Shopify, review your fulfillment SLAs at the SKU level.

7. Physical retail is holding: Forrester projects stores at 71% of total retail through 2030

The same Forrester report that projects $1.8 trillion in US eCommerce also projects 71% of all retail sales will still move through physical stores by 2030. Store traffic has proven more resilient than peak-pandemic forecasts suggested. Brands leaning on physical retail to build customer relationships while using eCommerce for fulfillment efficiency are increasingly the model that survives margin compression. (Retail Dive, Forrester)

The $20 Employee That Works While You Sleep

On January 16, 2026, Anthropic shipped something that should be on every merchant's radar. Claude Cowork launched as a research preview for Pro subscribers. Most people missed it. Yesterday, March 9, Microsoft announced Copilot Cowork, built directly on top of it in close collaboration with Anthropic. Microsoft's stock dropped 14% when the original Cowork launched in January. That tells you everything about what the software industry thinks this is.

Here is what it actually is: a desktop AI agent that runs on your computer, accesses the files and folders you give it permission to touch, and completes multi-step work autonomously while you do something else. Not a chatbot. Not a prompt tool. An agent that makes a plan, shows you the plan, executes it, and delivers finished files to your filesystem.

It costs $20 per month on Claude Pro. Available today on Mac and Windows. And it just became the most important productivity tool a mid-market merchant is not using.

What makes Cowork different from a chatbot

In a regular Claude conversation, you prompt, Claude responds, and you act on the response. Every step requires you. Cowork breaks that loop. You describe the outcome, give it folder access, and step away. Claude generates a plan, surfaces it for your review, then executes it step by step while you are in a meeting or asleep.

It can coordinate multiple workstreams in parallel. It supports scheduled recurring tasks, meaning you can tell it to run a report every Monday at 7 am, and it does. It integrates with Google Drive, Slack, Salesforce, and other tools via MCP integrations. Conversation history stays stored locally on your device, not on Anthropic's servers. That’s important if you are working with vendor contracts, margin data, or anything you would not want sitting in a cloud chat log.

Five workflows you should run in Cowork this month

1. Weekly sales digest, fully automated. Point Cowork at your Shopify export folder. Give it a report template. Schedule it to run every Monday morning. It reads the new CSV, populates the template, builds working formulas in Excel, and saves the file to your shared drive before you arrive at your desk. Zero manual work after the first setup. On a Team plan, the output lands in a shared Claude Project that your whole ops team can access.

2. Product catalog gap audit. Export your full product catalog. Ask Cowork to identify every SKU missing a meta description, weight, material attribute, or structured size data. Have it draft the missing copy using a brand voice document you provide as a reference. Output: a completed, import-ready CSV. What takes a merchandising coordinator two days takes Cowork under an hour. And it does not get bored halfway through SKU 400.

3. Vendor inbox triage and draft responses. Connect Cowork to Gmail or Outlook through the MCP connector. Ask it to review your inbox daily, flag anything from vendors or 3PLs requiring action, and draft replies grounded in your standard terms document. You review and send. It never sends without your approval. Most ops managers who try this recover 45 minutes a day within the first week.

4. Competitive pricing digest. Pair Cowork with Claude in Chrome. Give it a list of competitor URLs and your key SKUs. Schedule it to run weekly, pull current prices, compare against yours by category, and produce a one-page digest flagging where you are out of position. No manual spreadsheet. No agency retainer. $20 a month.

5. Post-season inventory debrief. After any major sale event, point Cowork at your inventory and orders export. Ask it to identify which SKUs sold faster than forecast, which underperformed, what the margin impact was by category, and what a revised buy plan looks like based on those patterns. Output: a formatted Excel file with a plain-English written summary. A task your analyst would spend a full day on.

Claude Projects: the memory layer that makes your team smarter

Cowork is the execution layer. Projects is the memory layer. They work together.

A Claude Project is a persistent workspace where you upload documents, set standing instructions, and have every conversation draw from that shared knowledge base without re-uploading anything. On Team and Enterprise plans, Projects are shared across your organization. Paid plans use RAG to handle large document sets, meaning you can load your full SOPs, vendor contracts, brand guidelines, and product catalog and Claude references them accurately in every conversation.

The merchant use case that pays for itself immediately: build a Project for your customer service team. Upload your return policy, shipping SLAs, product FAQs, and brand voice guide. Every rep now answers questions with your actual policy, not a generic approximation. Consistency across every ticket. New hires onboard against the same knowledge base from day one.

Build a second Project for your buying team. Upload supplier terms, historical PO data, and category performance reports. When you evaluate a new vendor, Claude reads your actual documents rather than giving you generic eCommerce advice. The comparison is grounded in your real numbers.

The signal the software industry has already read

Microsoft's stock fell 14% after Cowork launched because the enterprise software industry immediately understood what most merchants have not yet processed: an agent that executes knowledge work autonomously at $20 a month compresses the ROI case for a lot of SaaS tools and junior coordination hires simultaneously. That is why Microsoft built Copilot Cowork on top of it rather than competing with it.

That is not an argument to cut your team. It is an argument to stop having your team do work an agent can do, and redirect their time toward judgment calls, relationships, and creative decisions that actually require a human. The merchants who figure that out now, while Cowork is still in research preview and most competitors have not touched it, are building an operational advantage that compounds.

Download the Claude desktop app at claude.com/download. Switch to Cowork mode. Start with one workflow from the list above.

Shero News

Site Launch Spotlight

February and early March were busy. We launched multiple client sites, with Dillanos Coffee Roasters among the standouts. Dillanos is a specialty coffee roaster with a passionate customer base, and the team delivered a clean go-live. A full team effort from start to finish. More launches to announce in the coming weeks.

Celebrating the Women of Shero

March is International Women's Month, and we have 26 women on our team spanning multiple continents and time zones. They run projects, lead client relationships, manage engineering teams, and keep this agency operating at the level it does.

To every woman on our team and around the world: THANK YOU. This month and every month.

Meet Us at Shoptalk Vegas

Shero will be at Shoptalk in Las Vegas later this month. If you are attending and want to connect, reach out to me on LinkedIn. Always happy to talk Shopify, B2B commerce, or what is actually moving the needle for mid-market merchants right now.

Final Thoughts

To be honest, I am struggling to keep up with how fast AI is moving and how we can all use it effectively day to day. It is nerve-wracking at times. But it is also one of the greatest times to be alive.

The question is not whether to use it. It is how fast you can make it part of how you actually work.

A few things to act on before next month: migrate Scripts before June. Run the audit this week. Build the GA4 AI traffic segment today. None of these requires a big decision.

On that note, the next edition will include a proper guide to using Claude day to day as a merchant, with real workflows, clear examples, and ready-to-use prompts and skills you can apply immediately.