Working with Noibu in Claude

The Noibu plugin for Claude turns common ecommerce analysis questions into structured, repeatable workflows. It ships with a set of pre-built skills — packaged instructions to produce a specific kind of output, such as a ranked list of revenue opportunities, a checkout funnel chart, or the root cause of a priority error. Alongside the skills, you can ask Claude any ad-hoc question about your store and it will query Noibu directly.

This article covers the skills included in the plugin, with real merchant examples of how they're being used.

Ecommerce skills included in the plugin

Skills are useful when you want a deep, more structured analysis than a single question would produce, or when you want a result you can save, schedule, or share with your team. You can trigger a skill in three ways:

• Ask a question that matches the skill's purpose. Claude detects and runs the appropriate skill (for example, "where are people dropping off in checkout?" triggers the Checkout Analysis skill).

• Type the skill name as a slash command. For example, /find-opportunities or /store-pulse.

• Ask Claude to run a specific skill. For example, "run a full segment analysis."

Setup Noibu

What it does: The first thing a new user should run. It checks the connection status of the Noibu plugin, confirms you're authenticated, and verifies Claude can access data for your domain. 

When to use it: Right after installing the plugin, before your first analysis. Also run it any time you're unsure which domain or subdomain your data is coming from, after adding a new data source, or when a connection breaks.

How to trigger it: Type /setup-noibu to run this skill. 

Store Pulse

What it does: Gives you a live, at-a-glance overview of your store's health, presented as a Live Artifact that tracks your key ecommerce metrics. Use it to check overall store performance, customize what the dashboard tracks, or set up a recurring summary.

When to use it: First thing in the morning, before a leadership update, or any time you want a quick read on how the store is doing. Power users also use this skill to stand up a shared, always-current dashboard the rest of the team can open without asking a question.

How to trigger it: Ask "how is my store doing today?", "give me a store overview", or type /store-pulse.

What you get: A live dashboard artifact tracking conversion rate, revenue, sessions, and other core metrics for your selected time window. The artifact can be reopened later and refreshed against current data.

Find Opportunities

What it does: Generates a ranked list of revenue opportunities across acquisition, funnel, experience, and product — a small set of high-impact findings, each with an estimated monthly revenue impact. Use it when you want a prioritized "what to work on next" answer rather than a broad health check or a single-area deep dive.

When to use it: At the start of a planning cycle, when you don't know where to focus, or when you want a fresh read on what's changed week-over-week.

How to trigger it: Ask "where am I losing revenue?", "find me my biggest opportunities", or type /find-opportunities. You can also scope the run to a single area — for example, /find-opportunities for acquisition.

What you get: A set of ranked opportunity cards, each with:

• A specific gap (segment, baseline, magnitude)

• An estimated range of monthly revenue impact

• A trend label (worse than last week, better than last week, or same as last week)

• An Investigate button that hands the opportunity back to Claude for a follow-up conversation

Checkout Analysis

What it does: Analyzes checkout performance and health. Surfaces where shoppers drop off in the checkout funnel, what payment and delivery methods customers are using, why completion rates are low, what priority errors are hurting checkout, and how cart and order values benchmark against your norms.

When to use it: When you suspect checkout friction, after a checkout change you want to validate, or when conversion rate moves and you need to know why. A common pattern is a soft revenue day that prompts a funnel-depth check and an error check in the same conversation — this skill covers both.

How to trigger it: Ask "where are people dropping off in checkout?", "are there errors on my checkout pages?", "what payment methods are customers using?", or type /checkout-analysis.

What you get: A broad overview (full funnel by depth, cart and order value baseline, payment method mix, delivery method mix, and any priority errors on checkout pages) followed by targeted follow-ups on the most interesting signals. The final report leads with 3–5 key findings paired with recommended actions.

Segment Analysis

What it does: Analyzes segment and traffic performance. Surfaces which channels, devices, or countries convert best, how mobile compares to desktop, where your traffic is coming from, and which segments are underperforming.

When to use it: When you want to understand who's converting and who isn't, when channel performance shifts, when you're allocating budget across markets, or when you need to know whether a problem is universal or concentrated in one segment (for example, worse on mobile, or specific to one country).

How to trigger it: Ask "which channel converts best?", "how does mobile compare to desktop?", "which countries are underperforming?", or type /segment-analysis.

What you get: A broad overview across device, country, and marketing channel — each with a table and a specific insight callout — followed by targeted follow-ups on the most interesting signals (campaign breakdowns, landing page breakdowns, or funnel stage breakdowns by segment).

Product Analysis

What it does: Analyzes product and collection performance. Surfaces which products are underperforming, what your best-selling product type is, why a specific product isn't converting, how collections are performing, which products get views but no sales, and where shoppers drop off in the product funnel.

When to use it: Before merchandising decisions, after a product page redesign, when a specific product or collection feels off, or when planning a seasonal promotion and you need to find candidate products with strong customer interest.

How to trigger it: Ask "which products are underperforming?", "how is the Sale collection performing?", "why isn't my polo converting?", or type /product-analysis.

What you get: A broad overview across products, collections, and product types — views, add-to-carts, purchases, conversion rate, and revenue per session — followed by targeted follow-ups on the most interesting signals.

Tech Diagnosis

What it does: Investigates a specific technical issue or performance problem using Noibu data. Point it at a priority error, a symptom, or a slow page, and the skill summarizes the issue, traces the likely cause across the real sessions, errors, and performance data Noibu collects, and recommends a fix. It covers both technical issues (JavaScript errors, failed requests, checkout-blocking bugs) and page performance problems (Core Web Vitals).

When to use it: When a priority error is hurting checkout or conversion, when you need to confirm a fix actually landed after a code deploy, when a page is loading slowly or a layout-shift regression appeared after an app or theme change, or when you need to scope a technical issue clearly enough to escalate it to engineering. It is also the right skill for verifying a regression: run it before and after a deploy to see whether the error or performance metric improved in the data.

How to trigger it: Ask "what's the root cause of issue #299?", "is this checkout error still active after my deploy?", "which pages are slowest, and on which devices?", "did my CLS regression clear up?", or type /tech-diagnosis.

What you get: A structured diagnosis that walks through the issue summary, the supporting evidence (affected sessions, error detail, and performance metrics), the most likely cause, and a recommended fix. 

When you're ready to act, the skill can help you push a fix (GitHub connection required), package the diagnosis as a shareable report, or open a ticket in your engineering tool (for example, Jira or Linear) through Claude's own connector.

Saving, sharing, and scheduling

After an analysis runs, Claude offers a few ways to keep the result:

• Live dashboard. A persistent artifact in Cowork that you can reopen any time. The artifact re-runs the underlying queries each time you open it, so the data stays current.

• PDF snapshot. A static export of the report at the time it was generated. Good for filing or sharing externally.

• Slack or email. Some skills offer to send the result directly. For Find Opportunities specifically, you can also set up a recurring digest on a schedule of your choosing.

• No save. Leave the result in the conversation. You can always re-run the skill later.

Find Opportunities and Store Pulse support scheduled, recurring runs out of the box. For other skills, you can ask Claude to set up a custom scheduled task that calls them — for example, a daily store brief that pulls new priority errors, your conversion-rate pulse, and revenue versus the prior day, then delivers it to Slack, Notion, Google Drive, or your inbox each morning. This is a common way to get a structured "is anything wrong today?" summary pushed to you without opening a dashboard.

Other ways merchants use Noibu in Claude

Some of the most common merchant workflows aren't tied to a single pre-built skill — they combine ad-hoc queries with inline visualizations like session replay, click maps, and journey paths. A few patterns we see repeatedly:

Watching session replays to confirm a hypothesis

After identifying a friction point or error in the data, merchants often want to see what actually happened in a specific session — to confirm a hypothesis or explain the failure to the team. One men's designer apparel brand found a repeating dead-zone interaction pattern by watching high-rage-click session replays inline that aggregate data alone had missed.

Mapping click and tap behaviour to inform design decisions

Merchants combine click-text analysis, dead-click identification, and page-group journey shapes to ground redesign decisions in what shoppers actually do. A specialty bridal retailer mapped real mobile CTA click patterns to inform a homepage redesign, surfacing a dead zone that accounted for most above-fold interactions.

Measuring AI assistant referral traffic

A small but growing group of merchants is starting to treat AI assistant referrals (ChatGPT, Perplexity, Claude, Gemini, Copilot) as a distinct traffic channel. They track AI assistant referral conversion rate as the foundation of a generative engine optimization (GEO) strategy, using ad-hoc UTM and referrer queries against Noibu's session data.