Fortune 500 Omnichannel Integration
Native iOS front-end integration connecting customer-facing mobile app with in-store inventory and regional distribution centers for a $25B+ retailer.
Key Results
Technologies Used
This particular client is massive, with 1000s of developers worldwide. Sapling engineers were tasked with building the Native iOS front-end, while the client's backend developers would be responsible for creating the necessary APIs. The systems required specific knowledge of the business and limitations within the platforms themselves.
Challenges & Solutions
Velocity Mismatch Between Teams
While the App teams leveraged Agile with typical bi-weekly releases, the backend teams utilized a hybrid version of SAFe and waterfall that saw only one production release per quarter. If a backend dependency missed a Q1 deliverable, it wouldn't be available in production until late Q2.
How We Solved It
Our front-end developers collaborated with the client's QA and backend developers to create mock API responses. We shared those mocks with our backend counterparts so they had a contract to work towards. On the mobile side, we leveraged a cloud-based configuration switch to use either the mock or the real API call, allowing us to continue building independently and then 'flip the switch' once the API was ready.
Key Takeaway
“Don't 'wait' for an API - partner with your dependencies and drive towards a solution, and ultimately, the MVP.”
Legacy XML Format Instead of JSON
The client preferred to leverage XML rather than JSON. In the mobile world, XML is considered a bloated format, as JSON payloads are much smaller and more performant.
How We Solved It
We compared two API responses, one XML and one JSON, and shared documentation from our partners at Apple. Initially told it wasn't possible to use JSON due to platform limitations, one of our engineers dove into the documentation and showed the client how changing one build parameter would switch the API from XML to JSON.
Key Takeaway
“Know what tech stack your dependencies are leveraging, even if it's not your responsibility, because customer experience is paramount.”
Inconsistent API Behavior
One of the dependent APIs was consistently inconsistent - it would respond flawlessly for a dozen requests, and then inexplicably stop responding altogether for a few seconds before responding to API calls again.
How We Solved It
We spun up a Postman Runner to execute multiple, simultaneous requests to the API. We learned that two requests made at the exact same millisecond would cause server 500s. The root cause was Order Date stored in Unix Epoch Time as a primary key - submitting two orders in one millisecond caused the database to lock up. After the fix, we stress-tested at 150% anticipated capacity for days with no issues.
Key Takeaway
“We test our foundation so that our phone doesn't ring when a dependency goes down.”
“The Sapling team's ability to navigate our complex organizational structure while delivering high-quality work was exceptional. They didn't just build features - they helped us identify and solve systemic issues.”
Ready to get to the root of your problem?
Let us know your pain points by booking a complimentary consultation today. We come prepared to help.
Trusted by 50+ retail brands across the Midwest