Connecting a store to suppliers or B2B marketplaces
Connecting a store to one or more suppliers, or to a B2B marketplace, is a type of project I take on regularly, with variations depending on the quality of the API available.
The typical need
A store that relies on external suppliers needs its catalogue, stock and sometimes orders to stay in sync with what the supplier or marketplace actually exposes. This need shows up in several forms: importing a supplier catalogue for resale, publishing your own catalogue to a B2B marketplace, or doing both at once when a store buys on one side and resells on the other. In every case, the stakes are the same: keeping data accurate with no manual re-entry, without a failure on one side silently breaking the other.
How I approach this kind of work
I always start by studying what the partner's API actually exposes, not just what its documentation claims: call rate limits, fields that are genuinely reliable, undocumented behaviour discovered in practice. I then define a pivot data model on the store side that doesn't depend on the supplier's own format, so a supplier can be added or removed without rebuilding everything.
The synchronisation itself is built around one central question: who's right when the two systems disagree on the same piece of data. This priority rule needs deciding before development, not discovered at the first real conflict. I systematically add logging of the exchanges and, depending on how critical the data is, an alert on repeated failure, so a desync is spotted quickly rather than building up silently.
Factors that affect the estimate
-
Quality of the documentation
A well-documented API greatly reduces the time spent on preliminary study compared with one discovered by trial and error or by analysing network traffic.
-
Direction of the exchange
Syncing in one direction only is simpler than managing a two-way exchange where both systems can modify the same data.
-
Expected frequency
Real-time synchronisation calls for a different architecture from a batch sync run once or twice a day, with stricter reliability constraints.
-
Data volume
A catalogue of a few hundred SKUs is handled differently from one with several tens of thousands, particularly in processing time and partial-failure handling.
Frequently asked questions
Is an API required to integrate a supplier?
What happens if the partner changes their API without warning?
Can several suppliers be integrated in parallel?
How long does this type of project take?
Describe your need in one minute
A few targeted questions so I can reply with an estimate rather than another questionnaire.