At Just Eat for Business our customers are under particular pressure ordering meals for teams of busy colleagues who expect their food to arrive when they need it. As our orders grew we noticed a pattern of a high number of queries on orders that were on time and with no changes resulting in high effort for customers, taking time to call or chat to check up on their food, and also putting pressure on our Customer Care team.

I was personally responsible for the successful UX and design in a cross-functional team with stakeholders from Ops, Customer Care (CC), Product, Design, Data and Engineering. Together we agreed a key metric of Reducing inbound queries on orders that had no changes and that were delivered ‘on-time’ (within the first 10 minutes of the delivery window), that we felt directly reflected a successful reduction in effort for both customers and our Customer Care team.

From our kickoff the requirements from Ops and Customer Care were clear, they needed to reduce time spent checking the status of orders which were essentially on track. 
‍
Customers were sent a series of notifications by email and SMS that were triggered when orders were being prepared, out for delivery and then had been delivered.

I spoke to a range of ordering customers to better understand their experience between placing and receiving their orders and the affinities between them became clear.

• Uncertainty that notifications accurately informed them of the time of delivery
• Missing notifications in their email inboxes
• Feeling out of control having to wait to be notified about any potential issues
‍
Mapping our increasingly complicated notification algorithms revealed that conditions would always remain where they may not provide the confidence we hoped to achieve for customers. 

Several hypotheses were discussed to ‘fix’ the existing notifications, which I endorsed but following research around UX good practice in this area I proposed we develop additional self-serve status updates to put our customers in control of their updates.

I explored the status systems provided by equivalent B2C services and evaluated them against our differing requirements for customers who may order infrequently and so not be so familiar with our order management process and larger customers who may receive several orders a day from different restaurants. 
‍
I mapped the process of our various order types with the inputs we received from our system, restaurants and delivery partners and proposed several solutions that provided varying levels of granularity that we tested with interactive prototypes and weighed against cost and speed to market.

One of my early concepts included live point-on-map tracking for multiple-restaurant orders. As we ran through scenarios we realised that with delivery partners sometimes picking up and delivering several orders at a time that quite often drivers were shown following a route that users didn’t expect. This experience didn’t improve confidence for customers tracking their orders so despite initial enthusiasm was removed from designs.

We agreed on a solution that we could iterate which I provided designs for that could be built and released, evaluating engagement and customer behaviour at each stage.
‍
Though initial customer behaviour prioritised desktop usage we prioritised tight mobile accessibility as we hoped that we could drive engagement by directing our customers to these pages through our order management process and improved notifications.

I iterated designs from the most complex conditions back to the simplest, ensuring that consistent and useful feedback was provided across all scenarios using existing patterns. It was crucial that the progress of their orders were communicated succinctly with essential information available at each stage and changing available actions clearly visible.

With this iterative approach we were able to build and release the functionality for order types to maximise benefit and learning against engineering effort and adapted the implementation as we saw unexpected behaviour that we could accommodate.

Over the course of a number of releases, evaluations and optmisations we successfully reduced inbound queries on orders within our defined metric significantly. This was achieved through the work described, improving our notifications to highlight status updates, aligning language across channels and giving better visibility to the order management process in our customer accounts.

We were able to show good traction on page views for order tracking and inferred higher levels of self-service from this but would like to have better evaluated customer confidence and happiness as we improved their status tracking. 
‍
The long-term benefits to customers of providing both push and pull updates on their orders will continue to provide confidence and value for both customers and the business. It helped to raise the bar with expected functionality of our product creating further opportunities and reducing inbound queries generally as well as on orders with no issues. it is always worth looking at specific drivers of unbound queries however as we reduced inbounds by a similar amount by correcting unclear text on a sign-in error message that had been directing users to contact Customer Care, one small issue in an area of massive reach,