claude skills / E-commerce pack / audit-de-boutique / en français

Store Audit.

What changed, why, and three actions for this week, each with a name next to it.

See what it produces
  • Shopify
  • WooCommerce
  • PrestaShop
  • Google Analytics 4
  • Google Search Console
3 factorstraffic, conversion, average order, with each one's share of the gap
10actions at most, ranked by impact, confidence and effort
3actions for this week, with owner, date and metric

orders of magnitude, depending on your volume

the problem

What gets in the way today.

  • 01Sales are falling while traffic is rising, and nobody knows why.
  • 02The last audit ended with thirty recommendations, and none were done.
  • 03A new theme changed a detail in the checkout funnel, and nothing flagged it.

input, output

You give this. You get this.

The store exports and the context

Atelier Lisière, Shopify. October 2026 vs October 2025.
ventes_dans_le_temps.csv, sessions_par_source.csv, tunnel_par_appareil.csv, ventes_par_produit.csv.
Context: new theme installed on 2 September, sponsored Instagram campaign in October.
Store URL provided to check the pages on mobile.

$ claude "Run the audit: why are sales falling, and what should we do this week?"

Store audit: Atelier Lisièreexample

In short

Net sales fell by €536 (-4.5%) while traffic rose by 21%. The drop comes from the conversion rate, down from 1.60% to 1.20%, mostly on mobile between cart and checkout.

Sales breakdown

Sessions: 18,400 vs 15,200, +21.1%

Conversion rate: 1.20% vs 1.60%, -0.40 point

Average order value: €52.00 vs €49.50, +5.1%

Contributions: traffic +€2,247, conversion -€3,362, average order +€579. Total -€536.

Findings

1. Fact: on mobile, cart to checkout drops from 52% to 38% (tunnel_par_appareil.csv).

2. Inference: the new theme no longer shows the free shipping threshold on product pages, so shipping costs come as a surprise in the cart.

3. Fact: the Fig candle, the top product, was out of stock 12 days out of 31.

4. Hypothesis: the abandoned cart email is no longer sent. Test: place a test order and abandon the cart.

Actions for the week

1. Show the free shipping threshold on mobile product pages. Metric: mobile cart to checkout. Founder, 23/11.

2. Restock Fig, stock alert at 30 units. Metric: Fig sales. Workshop, 24/11.

3. Test then reactivate the abandoned cart email. Metric: recovered carts. Founder, 23/11.

Next check-in

Friday 27 November: mobile cart to checkout conversion, Fig sales, abandoned cart emails sent.

the method

What Claude does, in order.

  1. step 1Check the data

    Periods, rows, currency, totals that match. Test orders and outliers removed and named.

  2. step 2Break down sales

    Traffic, conversion, average order, with each one's contribution to the gap, calculation shown.

  3. step 3Go one level deeper

    Traffic sources, funnel steps by device, customers, products, key pages checked on mobile.

  4. step 4Rank the actions

    Each finding labeled fact, inference or hypothesis, then 10 scored actions and 3 for the week.

demo in French / click, try it

Your turn.

An interactive demo of the skill. It is in French, the pack works in any language: ask Claude to answer in English.

4 actions classées

constat

Sur mobile, le passage panier vers paiement tombe de 52 % à 38 %. Le nouveau thème n'affiche plus le seuil sur les fiches.

correction

Score 100. La fondatrice, le 23/11. Mesure : panier vers paiement sur mobile.

the film / 44 seconds, with sound

The pack film.

the pack film

In 8 seconds.

the guardrails

What the skill never does.

  • Only the store's data: a missing export is named along with where to find it.
  • No invented industry average, the store is compared with itself.
  • Every finding is labeled fact, inference or hypothesis, and every calculation is shown.
  • No promised numbers, and small volumes are flagged, not interpreted.

Questions

Which exports do I need?

Sales over time, sessions by source, funnel, sales by product and customers, over two periods of the same length. Failing that, the full order export as a CSV lets you recalculate almost everything.

Why no comparison with the industry average?

Because an unsourced average misleads more than it helps. The store is compared with itself, and an external benchmark is only used if you provide it with its source and date.

Does the skill change my store?

No. It analyzes and recommends. You decide and apply the actions.

Why only three actions this week?

So they actually get done and their effect stays measurable. The skill also checks that they don't all target the same page.

the pack

In the pack: E-commerce pack.