Margly
Getting started

First data load — when you'll see the first overview

What happens from clicking "Connect e-shop" to the first overview in the dashboard.

4 min read

What happens after the connection#

As soon as you connect an e-shop (Shoptet, Upgates, Shopify — add-ons for Shopify and Shoptet are in preparation) or an ad system or external service (Google Ads, Meta, Sklik, Heureka, Google Analytics 4, Search Console), Margly starts pulling historical data. It's not instant — it depends on the size of your e-shop.

What is pulled and in what order#

  1. Products (1–3 minutes)

    Margly first pulls the product catalog — names, SKUs, prices, purchase prices (if the e-shop provides them), categories, inventory. Without products you can't calculate margins.

  2. Customers (1–5 minutes)

    Customer database with emails, names, city, registration date. Used for LTV and repeat order calculations.

  3. Orders (5–30 minutes)

    Most of the data. Margly pulls the full history that the e-shop provides via API/feed. It takes orders, items, shipping, discounts, cancellation statuses. For ongoing updates, Margly then checks changes over the last 60 days with each load (typically catches cancellation, refund, status changes).

  4. Margin calculation

    After pulling, Margly recalculates margins, ABC analysis of products, customer cohorts. This takes another 1–2 minutes.

Ad systems and external services run in parallel#

As soon as you connect ad systems or external services available in the app, Margly pulls the data of these platforms in parallel with the e-shop. It's usually done within 5 minutes, because there is an order of magnitude less ad data.

How do I know loading is done#

  • In the Integrations section in the Data updates block you see the status of each platform. Green dot = done, yellow = running, red = error.
  • After completion you'll receive an email "First overview is ready" with a link to the dashboard.
  • Real numbers appear in the dashboard (until then you see placeholders).

What if loading fails#

The most common causes and what to do about them:

  • Invalid API key / token — go back to the Integrations section and renew the key. The procedure differs by platform:
    • Shoptet (API key) — in the Shoptet administration generate a new API key and paste it in Margly
    • Shoptet (CSV feeds) — in the Shoptet administration refresh the private URLs of feeds (orders, products, customers) and paste the new URLs in Margly
    • Shoptet (authorized add-on — coming soon) / Upgates (add-on) / Shopify (add-on — coming soon) — the token is passed to you automatically by the add-on, so in case of failure you need to reinstall the add-on in the given platform's marketplace
    • Ad systems and external services via OAuth (Google Ads / Meta / GA4 / Search Console) — click Reconnect and go through OAuth again with the correct account
    • Sklik / Heureka (API key) — in the given platform generate a new API token and paste it in Margly
  • Timeout / error on platform side — large e-shop or weaker API. Margly will retry at the next scheduled load. Frequency differs by tier — Starter 1×/day (morning), Professional 2×/day (morning + evening), Business / Enterprise / Trial 4×/day (morning, midday, evening, night). Details in Data sync. You'll learn about the failure as an in-app notification (max 1× per day per e-shop, aggregates all slots and data types — we don't send email in this case).
  • Limit exceeded (429) — the platform blocked you for a while (Google Ads has daily limits). Margly waits until the next scheduled load.
  • Permissions error — in Google Ads / Meta you don't have access to some account that Margly OAuth sent. Re-do OAuth with the correct account.

What's next after the first overview#

  1. Go through the Dashboard overview — what each card means.
  2. If purchase prices for products are missing (Margly reads them exclusively from the connected platform — Shoptet, Upgates, Shopify), fill them in your e-shop administration. Margly will update them on the next data load.
  3. After 7 days try the AI Advisor — it needs at least a week of data to get to know you.

Was this helpful?

Related articles