Mapping the landscape
Which systems are involved, such as the webshop, e-conomic, Business Central, CRM and warehouse, and what is keyed in more than once today.
On paper it sounds simple: data has to travel from one system to another. The catch is that customers, item numbers and units are rarely recorded the same way, and software cannot decide which system is right. That call has to be made by the business.
The technical side can almost always be solved. The real work is reaching internal agreement on which system leads for each kind of data.
Which systems are involved, such as the webshop, e-conomic, Business Central, CRM and warehouse, and what is keyed in more than once today.
What goes where, how often, and which system owns customers, products, prices and orders.
We link item numbers, customer records, VAT codes and units across the systems.
Through the official APIs, or through a middleware layer we build where no API exists.
What happens when a system is down: retries, queues and an alert to a person when a decision is required.
We keep an eye on the runs and notify the owner if an integration stalls.
A simple integration between two systems takes about two weeks. Linking several systems with shared master data takes six weeks or more.
We review master data at both ends. Duplicates and inconsistencies usually surface and must be cleaned up before we start.
Data flow, rules on which system wins and how conflicts are resolved.
Built and tested in a staging environment with real but pseudonymised data rather than invented samples.
A period running alongside the manual process, after which manual entry stops.
Data edited in two systems at once is the biggest source of conflict. A salesperson updates a customer's phone number in the CRM, the bookkeeper does the same in the finance system, and the integration overwrites one with the other. That is why a leading system is appointed for each data type in advance, and the field is locked in the other.
Clean it up before integrating, and that is a job for whoever owns the data, not for the developer. Connect two databases full of duplicates and discrepancies and the mess simply spreads to both.
Rarely. Most tasks are well served by hourly runs or a few runs a day, which is also more robust. Real time is needed where decisions are made on the spot, such as checking stock at the moment a customer places an order.
There are workarounds: file exchange, direct table access or automating the user's clicks. Each is inferior to an official API and needs care when the software is upgraded, but the goal is usually reached.
Yes, for example e-invoicing via NemHandel in OIOUBL format and CVR lookups. Integrations handling personal data are documented so you can account for them to Datatilsynet.
Tell us which systems your staff currently copy data between by hand. We will estimate the work and tell you what needs tidying first.
Thank you for getting in touch
One of our consultants already has it. Expect a reply within the working day; anything urgent goes straight to an engineer.
That city is not on our list. Check the spelling or pick the nearest larger town.