Someone is doing the same update twice
A parcel goes out in the shipping tool and the retail system has to be told by hand. Whatever the pair of systems, that second entry is the job we automate, triggered by the action staff already take.
Connect the work behind the storefront
Orders, stock and shipments move through several systems, and someone is usually filling the gaps by hand. We build the connections, with clear rules for what moves, who owns it and what happens when it doesn't match, so the second entry stops being a person's job.
An order doesn't end at checkout. It has to reach the warehouse, come back as a shipment, sometimes come back again as a return or an exchange, and turn up in the accounts. Every one of those handovers is a place where two systems can disagree about what just happened, and where a customer or a buyer tends to find out before you do.
We build and maintain the connections between ecommerce, retail, warehouse, wholesale and finance systems: Shopify, Datapel, Apparel21, Retail Express, Magenta, Cin7, NuORDER, ShipStation and Starshipit so far. The work starts with a real transaction and the person handling it, then gets into identifiers, timing, ownership and the exceptions. What you end up with is a flow you can explain to a new staff member, plus the checks and recovery steps for when it needs attention.
Sounds familiar?
Usually one of these three. Often two at once.
A parcel goes out in the shipping tool and the retail system has to be told by hand. Whatever the pair of systems, that second entry is the job we automate, triggered by the action staff already take.
Regional pricing, product families, delivery windows, what an exchange does to stock. We work with your team to establish which system should own each rule and put the transformation where it belongs.
An API is changing, a platform is being replaced, or nobody can explain what the middleware does. We map the current flow, then repair, extend or replace it in stages you can verify. Sometimes a scheduled Matrixify import is enough, and we'll recommend it when it fits.
What you get
Which records move, in which direction, which system is authoritative and which identifiers tie them together. Field meanings as well as formats: variants, locations, statuses, timing.
Products, availability, orders, fulfilment, and where you need them, returns and exchanges. Each flow has one job. A shipment update and an order import can share a reference and still be different things.
Missing fields, ambiguous identifiers, a service that’s down, an event that arrives twice. We agree the behaviour up front, and records that need a human are held with enough context to decide.
Run summaries and reconciliation that compare what was expected with what arrived, and a handover that explains how to review and recover. A successful API response is not the same as a correct record.
How it goes
We use representative records to test the flow and its failure cases, then we agree a rollout in stages.
Representative records, the decisions around them, who owns what, and the exceptions your team already knows about. Access gets sorted here too.
Missing data, repeated events, changed records, odd statuses. We test these in a reviewable environment and compare the results with what the receiving team needs.
We agree the records and size of each rollout stage, then check the receiving system before expanding it. We investigate discrepancies and document the operating steps and who owns each part.
Recent integration work
One replaced existing middleware flows. One closes the dispatch loop.

ELK · Connected retail operations
Shopify, NuORDER, Datapel and ShipStation, connected around the sale, stock and shipping handovers.


Bicycle Superstore · Fulfilment integration
Dispatch information passes from Starshipit through Shopify to the matching Retail Express order.
Common questions
Shopify, Datapel WMS, Apparel21, Retail Express, Magenta, Cin7, NuORDER, ShipStation and Starshipit, all in published work. Whether a new pair can be connected depends on their APIs and your access, which we assess early.
Yes. We can investigate a specific problem, map the current flow or assess an extension. Then you'll know whether it needs a repair, ongoing maintenance or replacing.
No. Some changes need to be prompt, some can run on a schedule. We agree which is which and make sure operators can tell how current the information is.
Stable identifiers, a record of what has been processed, and retries that know about both. The exact controls depend on the APIs; repeated events are part of the testing.
The systems involved, a redacted or fictional example, and the outcome you need. Leave out customer records and credentials; we can agree how to review further detail.
Work directly with Chris
The systems involved, a redacted or fictional example, and what you want to happen.