BaseLinker integration from the order to margin after returns
BaseLinker collects orders from Allegro, Amazon, eBay, Shopify and WooCommerce into one panel. We connect it to whatever comes next: accounting, the CRM, the warehouse and the data warehouse, so the invoice, the label and the stock figure are produced without anyone retyping them. On top of that we add an AI agent that answers, from those same numbers, how much a single product really earns after returns.
- every 5 minuteshow often the automation checks BaseLinker for new orders, so the invoice and the shipment do not wait for someone to get back to their desk
- 2 to 6 weeksfrom the workshop to a working integration, depending on how many channels you sell through and how many systems sit on the other side
- EUR 3,500 netper week of the process workshop, price published up front; the rollout is quoted only once the workshop is done
The problem
Why does BaseLinker tidy up your orders and leave margin a mystery?
BaseLinker does well what it was built to do: it pulls orders from several channels into one panel and gives them a shared flow. The trouble starts at its edges. The invoice is created in another system, the cost of goods sits in accounting, returns live in a spreadsheet and ad spend is in the Google and Meta panels. Each of those places tells the truth about its own fragment and none of them tells the whole one.
That is why the statuses are where most of the manual work happens. Someone copies a tracking number into the message to the customer and into the CRM. Someone remembers, or does not remember, that an accepted return needs a credit note. Once a month someone lines up sales against the cost of goods to answer the question of which products are still worth advertising.
Part of that work stopped being a matter of convenience. Structured e-invoicing is becoming mandatory market by market across the EU, and every mandate turns the sales invoice into a document with a defined format and a defined moment of issue. An invoice a person still builds from an order by hand now carries one step more than it did a year ago, and that step is easy to lose on a busy Monday.
- The tracking number pasted separately into the email and into the CRM, on every single order
- Credit notes issued in a batch at month end, because nobody tracks them as they happen
- The same unit sold in two channels, because the stock update went out a minute too late
- Margin per product calculated in a spreadsheet once a month, which is too late for a budget decision
Why now
What will BaseLinker do on its own, and what needs an orchestrator?
BaseLinker has its own automatic actions tied to the order status. A status change can message the customer, issue an invoice, create a shipment with the carrier or move the order along. If there is one condition and every piece of data it needs already sits in BaseLinker, that is enough, and we will tell you so in the consultation.
An external orchestrator only earns its place when the decision depends on something outside the panel: stock in the ERP, a customer's credit limit in the CRM, that customer's return history, the exchange rate on the day of sale. Then we move the logic into n8n, Make or a direct connection through the interface, and BaseLinker stays the place where the order lives and holds one status. The technical layer, the methods and the rate limits, we cover separately under API integration.
The third change is the AI agent. A rule will move data from system to system, but it will not tell you which pattern comes back most often from Germany and whether it still earns anything after returns. BOND, our company AI agent, reads sales, returns, cost of goods and ad spend at once and answers in a plain sentence in 5 to 10 seconds, with nobody logging into four panels.
What changes
What does order handling get, and what does the owner get?
An order from any channel enters the same flow
The automation pulls new orders every 5 minutes and passes them on: to the invoicing system, to the CRM and to the data warehouse. The sales channel becomes a field on the order rather than a separate exception to handle.
One rule covering Allegro, Amazon, eBay, Shopify and WooCommerce at once
A status does more than change a colour in the panel
A status change fires a scenario outside BaseLinker: the message with the tracking number, the record update in the CRM, the event written to the warehouse. We set the order and the conditions in the workshop, because those are what decide which parts are safe to automate.
One event instead of three manual steps after a status change
A reservation takes the unit off stock before the second channel sells it
A paid order reserves stock, and the new figure reaches the shop and the marketplaces within seconds. We agree on one source of truth for stock, BaseLinker or your ERP, and every channel reads it instead of reading each other.
Stock counted in one place and broadcast, not retyped in every panel
The invoice and the credit note come out of the order data
Fakturownia, iFirma, Subiekt or Comarch Optima receive the data straight from the order, the invoice goes into the national e-invoicing system where you are required to file it, and the number comes back to BaseLinker. An accepted return triggers the credit note and a tag on the customer record in the CRM.
The credit note starts on the day the return is accepted, not at month end
The label and the tracking number with nothing copied
The waybill is built from the order data at the carrier connected inside BaseLinker, and the number goes back to the order, into the message to the customer and into the CRM. A status change at the courier updates the order, so support stops checking the carrier panel.
The tracking number in three places from one event
Margin per product, counted after returns
Sales, cost of goods, delivery cost and returns are pulled out of BaseLinker into <a href="https://cloud.google.com/bigquery/docs" rel="noopener">BigQuery</a> or a spreadsheet, lined up per product and per market. The AI agent answers from that data in a plain sentence instead of pointing you at another export.
One figure per product, refreshed daily instead of once a month
Proof
Do you run on BaseLinker yourselves?
Yes, in the shop one of our founders built. It is the only proof we can show without asking a client for permission to publish their numbers. Below is half a year of measurement, together with what that number does NOT prove.
Built for ourselves. These are not client deployments but implementations inside the founders' own companies. We say so plainly, because proof you cannot verify is not proof.
SurfBottheir agentSurf Inc., a premium clothing brand with an online shop and two physical stores
- 16
- people from the company work with the agent
- 21
- systems wired in permanently
- 89
- processes run without a human
- 323
- thousand lines of the agent’s code
Orders from several channels come into BaseLinker, and from there the data moves on to the ERP, the shop, the data warehouse and the ad platforms. The AI agent launched in March 2026 and was fully running by May. Sixteen people use it, the warehouse and the store staff included.
8 of 21 systems it works in
PrestaShop- BaseLinker
- Comarch Optima
BigQuery
Google Ads
Meta Ads
GA4
Zalando
- Before
- From February to August 2025 every unit of ad spend came back as 3.53 in online revenue, and advertising absorbed 28.3 percent of that revenue. Sales and returns per product were lined up by a person, in a spreadsheet.
- After
- In the same months of 2026 the ratio is 4.19 and advertising is down to 23.9 percent of revenue. The ad budget is 6.9 percent lower with sales 10.3 percent higher.
Return on ad spend from 3.53 to 4.19, 18.5 percent higher, on a budget 6.9 percent lower
And in your company
What if our BaseLinker is set up differently?
Four sentences we hear most often about this tool, and what we answer to each of them.
BaseLinker has automatic actions, why add another tool?
If there is one condition and the data sits in the panel, there is no reason to, and we will tell you that. We then show you how to set the actions up on the statuses and the whole thing ends at the workshop. An orchestrator starts to make sense when the decision depends on stock in the ERP, a credit limit in the CRM or a customer's return history.
We have BaseLinker and Comarch Optima, but nobody ever connected them.
That is the most common starting point. We begin with one direction, usually the sales invoice created from the order, and run it in parallel with your current process. Credit notes after returns and the cost stream in the other direction come only after that first flow checks out.
We sell on Allegro and Amazon, and each has its own return rules.
The mechanics of a marketplace return belong to the platform and no integration gets around them. So we work on what stays on your side: the credit note, the tag in the CRM, putting the unit back on stock, and the data on which product and which size comes back from which market.
We have no IT team and nobody here knows the BaseLinker interface.
They do not need to. What stays on your side is generating a token in the panel and deciding which data may be sent onward. The methods, the rate limits and the way authentication works we cover separately under API integration, and after the rollout we hand over the documentation and train someone in support to read the error queue themselves.
- Sales and marketplaces
- BaseLinker, Allegro, Amazon, eBay, Shopify, WooCommerce, PrestaShop
- Accounting and invoicing
- Fakturownia, iFirma, Subiekt, Comarch Optima, e-invoicing filing
- CRM
- HubSpot, Pipedrive, Salesforce, LiveSpace
- Shipping
- carriers connected inside BaseLinker, labels, parcel tracking
- Data and reporting
- BigQuery, Google Sheets, GA4, Looker Studio
- Automation layer
- n8n, Make, Zapier, custom API integrations
Events in BaseLinker and what follows from them outside the panel
Six events we usually start from. We set the order and the conditions in the workshop, because they decide what is safe to automate straight away and what is better run in parallel first.
| Event in BaseLinker | What happens next | Where the data goes |
|---|---|---|
| New order | sales invoice, contact and company in the CRM, a row in the data warehouse | invoicing system, CRM, BigQuery |
| Status change | message to the customer, record update, event written down | email, CRM, data warehouse |
| Shipment created | the waybill number goes back to the order and into the message | BaseLinker, email, CRM |
| Stock movement | post-reservation stock broadcast to the sales channels | shop, Allegro, Amazon |
| Return accepted | credit note, tag on the customer, the unit back on stock | invoicing system, CRM, warehouse |
| End of day | sales, cost of goods, delivery cost and returns per product | BigQuery, spreadsheet, dashboard |
A first rollout usually covers two or three of these events. The rest go in down the same track once the first ones pass the 30 day measurement.
How we do it
Five steps from the BaseLinker panel to a connected flow
Every step ends with something you can judge before you decide on the next one. We do not install tools blind.
Workshop
One week. We walk one of your orders from the moment it lands in BaseLinker to the moment a return is closed, and note every place where data is retyped by hand. We check how many channels you handle, which invoicing system and ERP stand on the other side, and which automatic actions in the panel already work.
Integration design
The list of events that should leave the panel, the direction of each stream and one agreed source of truth for stock. Then the tooling choice: automatic actions in BaseLinker alone, n8n, Make or a direct connection through the interface. Always the simplest route to the goal, not the most elaborate one.
Pilot on real data
The flow runs in parallel with your current process, with customer messaging switched off. We replay situations from your last season: two sales of the last unit in the same minute, a return filed after 14 days, an order with an empty tax number, a parcel lost at the courier.
Production and training
Once it checks out we switch to production with error monitoring. An off switch stays on your side, along with an error queue that shows what did not go through and why. Someone in support learns to read it and close it, without raising a ticket with us.
Measurement after 30 and 90 days
We check the metrics written down in the workshop: how many orders went through without a person, how many landed in exceptions, how many credit notes went out late. Only then do we add the next event, down the same track.
Before and after
How the day changes for the person who runs BaseLinker
Before
- The morning starts by going through the overnight orders and issuing invoices one by one, by hand
- Tracking numbers get pasted into the messages to customers and into the CRM
- Returns are logged in a spreadsheet and credit notes go out in a batch at month end
- Asked about the margin on a product, the answer is that a report will follow later
After
- Overnight orders have an invoice, a contact in the CRM and a row in the warehouse before the laptop is open
- The tracking number is with the customer and in the CRM from one event
- The credit note starts on the day the return is accepted, together with the tag on the customer
- Asked about the margin, the AI agent answers in 5 to 10 seconds
The first event is usually live in production 14 days after the workshop. The effect on handling time and on margin gets judged after 30 and after 90 days, on metrics written down before the start.
For the person signing this off
What the shop owner gets out of it
You stop depending on one person's memory. The rules on the statuses are written down and behave the same way when order handling is on holiday and when that person leaves. The sales invoice comes out of the order, so your e-invoicing obligation stops hanging on whether somebody remembered an extra step.
The second thing is the budget conversation. When sales, cost of goods and returns per product sit in one place, you know which products are still worth advertising and which ones eat budget and come back. You start with one event, so the first decision is small and reversible.
- The workshop ends with a document you can execute without us
- The pilot runs in parallel, so your current process works until the tests are over
- The off switch and the error queue stay on your side
- Tool costs go straight to the vendor, with no margin from us
What it costs
The workshop has a published price, the integration is quoted after it
The process workshop costs EUR 3,500 net per week and runs from 1 to 7 weeks. For BaseLinker one week is usually enough, two when several channels and a separate ERP come into the flow. We quote the integration after the workshop, because only then is it clear how many events leave the panel and what state the systems on the other side are in.
Your BaseLinker subscription, the cost of scenarios in Make or n8n and the cost of running the language model go straight to those vendors, at their list prices, currently EUR 0.005 to 0.07 per query. We add no margin and no invoices pass through us.
Process workshop
EUR 3,500 net
per week, usually one or two for BaseLinker
A self-contained piece of work. It ends with a map of the order flow and a list of events to connect, in an order you can execute without us.
- A map of the order path from the sales channel to a closed return
- The list of events to connect, with the direction of each data stream
- One agreed source of truth for stock
- A tooling recommendation: actions in BaseLinker, n8n, Make or the API
- The metrics to measure at 30 and 90 days
- Most popular
Workshop and BaseLinker integration
Quoted after the workshop
scope follows from the number of channels and systems on the other side
The option most shops choose. A working flow 2 to 6 weeks after the workshop, launched first in parallel with your current process.
- Everything in the workshop
- Connected to accounting and e-invoicing, to the CRM and to the sales channels
- Reservations and stock broadcast from one source of truth
- Labels, tracking numbers and carrier statuses
- Pilot on real data, error monitoring, training for your team
BaseLinker and an AI agent
Quoted after the workshop
for shops that ask about margin and returns every day
Everything in the integration plus BOND, which answers from that data in a plain sentence. Rolling out the agent takes around two months.
- Everything in the BaseLinker integration
- A data warehouse with sales, cost of goods and returns per product
- The AI agent wired into BaseLinker, accounting and the warehouse
- Answers in 5 to 10 seconds, in English and in Polish
- Access for support, the warehouse and the owner
The workshop ends with a document that stands up without us: the flow map, the list of events and the order of the rollout. If it turns out that the automatic actions inside BaseLinker are enough, we say so plainly and show you how to set them up. The matter then ends at the workshop rather than at an integration you do not need.
Frequently asked questions
What is a BaseLinker integration with other systems?
It is connecting the BaseLinker panel to what sits on both sides of it: the sales channels, the invoicing system, the CRM, the warehouse and the data warehouse. New orders and status changes leave the panel automatically, and the results come back to the order. BaseLinker stays the place where an order holds one status, while decisions that depend on data from other systems are made by the automation layer.
How long does a BaseLinker integration take?
One week of workshop and 2 to 6 weeks of rollout, depending on how many channels you sell through and how many systems come in on the other side. The first event, usually the sales invoice created from the order, runs in production around 14 days after the workshop. The next ones go in after the measurement, not all at once.
How is an automatic action in BaseLinker different from a scenario in n8n or Make?
An automatic action works on data that is already in BaseLinker and fires on a status change. It is enough when there is one condition. A scenario in n8n or Make earns its place when the decision depends on something outside the panel: stock in the ERP, a customer's credit limit, their return history or the exchange rate on the day of sale. The choice is made in the workshop and sometimes it lands on the actions in the panel.
Will the integration handle invoices and e-invoicing?
Yes, and that is the most common first stream. Data from the order goes to Fakturownia, iFirma, Subiekt or Comarch Optima, the invoice is filed in the national e-invoicing system that applies to you, and the number comes back to the order in BaseLinker. An accepted return triggers a credit note. If your invoicing system already files on its own, we do not duplicate that; we take care of what happens before and after.
How does stock sync work when we sell across several channels?
In the workshop we agree on one source of truth for stock: BaseLinker or your ERP. A paid order reserves the unit at the source, and the new figure reaches the shop and the marketplaces within seconds. Every channel reads the same number instead of reading each other, so the last unit does not sell twice. We cover the stock mechanism itself on the page about automatic inventory updates.
What about couriers and labels?
The waybill is built from the order data at the carrier connected inside BaseLinker, and the number goes back to the order, into the message to the customer and into the CRM. A status change at the courier updates the order, so support does not open the carrier panel to answer a question about a parcel.
Where does the data for margin after returns come from?
Sales, delivery cost and returns come out of BaseLinker, cost of goods from accounting or the ERP, and ad spend from the campaign panels. It all lands in BigQuery or in a spreadsheet, lined up per product and per market. That way margin is counted after returns rather than before them, and you see it daily instead of after the month closes.
Do we need a developer and access to the BaseLinker interface?
A developer, no. What stays on your side is generating a token in the panel and deciding which data may be sent onward. The methods, the rate limits and the way authentication works we cover on a separate page about the interface. After the rollout we hand over the documentation and train someone in support to read the error queue and add simple conditions themselves.
Let us walk the path of one of your orders together
The free consultation takes 30 minutes. You tell us how many channels you sell through, what sits on the other side of BaseLinker and which step a person still does by hand most often. We show you what can be connected straight away and what needs the data tidied up first. You do not have to share any system or any token at this point.
We run at most two rollouts at a time, so that someone is actually sitting on each of them.
Book a free consultationPS. If somebody in your company pasted a tracking number into an email and into the CRM today, that is exactly this problem. You write to Paweł, one of the founders, not to a shared inbox, and you get an answer within one working day.