15 mins read
04.
01.
+90%
01.
Improved ease of use
EXTREME MAKEOVER
INTERNAL LOYALTY TOOL EDITION
HIGHLIGHTS
A SCALED-BACK ENHANCEMENT
STREAMLINED, OPTIMISED, ENGAGING
01.
SIMPLIFIED CUSTOMER DASHBOARD
Unlike its predecessor, LAT only shows imperative customer info needed to conduct Loyalty admin tasks - designed with a focus on visual hierarchy, chunking and Miller's Law.
02.
NEW FEATURES, EVEN FOR THE MVP
There were a host of quick wins that we achieved for the MVP that have exponentially helped our users ie. Transaction History filtering.
With potential years of data to scroll through daily for users, they need a quick (and modern) way of expediting their process.
03.
ALL TASKS OPTIMISED
One of the main featues, Adjusting a Wallet, used to be split over six different screens. It's now condensed onto one form.
The architecture, and all tasks, have been tested rigorously to be shortened and enhanced.
04.
ERROR PREVENTION
The interface will now gently nudge you to double-check the before and after when conducting an admin task.
This will help stop accidental human error, and save unnecessary added stress of removing 100 Rewards from a customer's account before they find out (a true story).
PROJECT STRUCTURE
01
DISCOVER
User interviews
Contextual inquiries
Card sorting
Heuristic analysis
Workshop facilitation
02
DEFINE
Affinity mapping
Research synthesis
Feature prioritisation
Success measures
MVP scoping
03
DEVELOP
IA maps
User/task flows
Design iterations
Feedback sessions
Usability tests
04
DELIVER
Pre-prod testing
UAT
SUS survey
Tutorial guides / FAQs
Presentations
05
MEASURE
Bug fixes
Feature improvements
Post-MVP features
Collating feedback
Monitoring numbers
01 DISCOVER
"You cannot understand good design if you do not understand people. Design is made for people." - Dieter Rams
I did not choose the name for the Loyalty Admin Tool (LAT). I did petition to get the name changed to 'Luke's Admin Tool', but for some reason no one signed it.
project context
Nando’s have just completed a major two-year migration and are saying goodbye to their current loyalty provider, Paytronix, or as it’s known internally Pain-tronix. Paytronix (PX) is a legacy loyalty platform that turned even basic tasks into metaphorical marathons. It was slow, clunky, painful to navigate, and ran entirely on US time (that is not a bad joke).
We’re moving to Talon.One, a faster, more flexible loyalty engine that sets us up to grow our programme in the years ahead, while saving the business £500k+ per year.
WHAT'S THE PROBLEM?
In order to complete the migration, I needed to design the replacement for Paytronix - Enter the Loyalty Admin Tool (LAT). For the MVP, LAT will be a scaled-back version of Paytronix, but in order to achieve this I need to:
01
WHO
Find out how many users
Paytronix has and which
teams use the tool
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
02
WHY
Find out why users need
and use Paytronix for
their roles
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
03
TASKS
Collate all known tasks
users conduct in PX
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
04
STRATEGISE
Determine which tasks
are imperative to build for
MVP
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
05
ENGAGE
Create a tool that isn't
filling users with dread
every time they log on
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
Contextual inquiries
WHAT DOES PAYTRONIX LOOK LIKE?
HEURISTIC ANALYSIS
I'll keep this heuristic analysis very brief as there is a lot more than is shown here. I'll show you the 'customer look up' to 'adjusting a wallet' flow. This has been simplified for speed purposes, but have a look for yourself and try to figure out from the screenshots how you achieve this.
Spoiler alert: it took me far longer than it should have.
WHO ARE THE USERS?
01. I found over 200 emails tied to a Paytronix account.
02. I then sent a mass email asking them to share how and why they use PX.
03. Received 67 response from active users, and then grouped them by team.
04. Set up user interviews with two members from each team, 14 users in total
USER INTERVIEWS
The interviews were split up into two sections:
01. Contextual inquiry - seeing how our users conduct tasks in PX in their normal work setting.
02. Card sorting - I collated all of the known tasks from the mass email I sent and users moved tasks into columns of Not Important to Very Important for their roles.
This will be imperative to understand which features we need to rebuild, and design in LAT.
SYSTEM USABILITY SCALE
I gave the testers a tiny SUS survey post-interview to gauge how they view using Patryonix. This will help me benchmark LAT against PX, and can hopefully gain me some bragging rights.
Warning, the results may shock you.
SUS RESULTS
Paytronix officially scored a "Crikey" with a 39 score overall on my SUS scale. And to think these people have to use it every single day for their roles. Let's bump this up to at least an 85.
USER PAIN POINTS
A bonus finding was that the two main user groups of Paytronix, Customer Support and Safety & QA teams, were over the moon that their prayers of having an improved tool for their daily work were finally being answered.
02 DEFINE
"You can use an eraser on the drafting table or a sledgehammer on the construction site." - Frank Lloyd Wright
AFFINITY MAPPING
After conducting the heuristic analysis I had a vague idea at which areas could be improved upon in LAT. Having the users go into detail during the contextual inquiries and voice their frustrations was extremely elucidating.
Map was created in Condens and each white square contains the annotated video from the interviews.
RESEARCH INSIGHTS
01. FEATURE CONFUSION
A staggering amount of features were neither being used, or weren't understood. We'll have to scale it back for the MVP.
02. POOR NAVIGATION AND LEARNABILITY
The only way new users understood how to do certain tasks was to be shown, which doesn't help when the tool is so vast. Because of this, users were having to create workarounds for certain tasks, which disrupted the long-term data.
03. UI OVERHAUL
This one was a no-brainer, as you've seen the pictures at the start. Not just aesthetically, but functionally as well. Also, it's not the sexiest internal tool that conducts loyalty admin I've seen (wink).
04. USERS CLAIM TASKS "CLUNKY"
This is due to a host of different reasons, from the interface to the navigation, and also having to chop and change from a few different internal tools to juggle customer queries.
05. REDUCING COGNITIVE OVERLOAD
As you can see from the above images, there is far too much information displayed on nearly every screen - which doesn't aid comprehension or learnability at all.
SUCCESS METRICS
The success metrics effectively wrote themselves through the discovery and talking to the users, but we wanted to focus on these main four - which if we can make the users jobs slightly easier by achieving better scoring on these I'll be a happy camper.
MVP FEATURE PRIORITISATION
CARD SORTING RESULTS
Below you will see a disgusting excel table that I made, to which I sincerely apologise, to collate all of the card sorting findings. This was imperative to figure out from all users which features are critical for the MVP of the admin tool.
Remember, users need this for their day to day work, so we cannot leave them hanging by missing out a critical feature. No pressure, Luke.
It is also here that my PM said "I am the first designer he's ever worked with that has voluntarily used Excel". Luckily he doesn't know that I was crying silently whilst manually building that table.
03 DEVELOP
"There's no such thing as a wrong idea. There are only ideas we haven't found the right place for yet." - Rick Rubin
Information Architecture
Now came the fun part, beginning the backbone of the admin tool - the information architecture. This was a necessity for both myself and the engineers to not only see all of the must-have tasks in one place for the MVP, but how they might be able to be grouped together.
All of these tasks are numbered to match the tasks in the hideous Excel table above, and once opened have annotations and user feedback for each.
LAT USER FLOW
After collaborating with the engineers from both the card sorting results, and now the IA, we started working on the site map. The IA allowed us to prioritise features for MVP, and which features or functionalities can be grouped together.
Aforementioned, LAT will be only used for loyalty admin for customers on their Nando's Rewards account. The main tasks at a high level will be:
01. ADJUSTING A WALLET
02. CUSTOMER TRANSACTION HISTORY
03. editing card statuses
04. CENTRALISING CUSTOMER DATA
Adding/removing Chillies and Rewards to a Nando's Card - the main functionality for LAT.
Helping to validate restaurant visits, and Loyalty orders - critical for adjusting wallets.
Temporarily blocking a Nando's Card, and transferring card balance to a new card.
Having a one-stop-shop for reviewable customer data for Loyalty, Nando's ID, and Orders information.
The red shapes are where errors are likely to occur (thank you engineers for answering a lot of my dumb questions).
Side note: I would have loved to create a site map for Paytronix, but I would probably still be trying to create it nearly two years after this project started.
UX LAWS
JON YABLONSKI
There were a few core principles I wanted to focus on whilst designing the Loyalty Admin Tool, and my Christmas present of the Laws of UX cards by Jon Yablonski came in very handy. Thank you, Mum (Santa).
INITIAL CONCEPTS
{ PAYTRONIX }
{ LAT }
{ PAYTRONIX }
{ LAT }
{ PAYTRONIX }
{ PAYTRONIX }
{ LAT }
THINKING BEHIND THE DESIGN
EMPATHETIC SOLUTIONS
01. CUSTOMER SEARCH made easy
Customer search is extremely convoluted, and very overwhelming on PX. There are too many options displayed immediately and no sense of visual hierarchy.
I've removed the amount of input fields to drastically reduce cognitive load and improve simplicity. There is a slight technical constraint with the improved search in that each search entry is pulled from a different data lake - meaning that the user has to state which search type they are conducting. Enter the filter lozenges.
02. DASHBOARD UTILISATION
Currently upon entry to PX you are hit with the first image - a wall of helper text detailing what each section of the sidebar navigation does… Let's not.
I suggested during the heuristic analysis that utilising this space using links would be very beneficial for a host of reasons. What I've suggested here is to have a 'Your workspaces' section (the users can determine these during testing) and a 'Recently viewed' section for quick re-entry into previous tasks/journeys.
03. SIMPLIFIED NAVIGATION
In the 'Customer Support' section of the navigation, there are 22 links alone. I would have liked to count all of the links in each category but I don't think I can count that high.
Conducting the card sorting was critical for narrowing the scope of this tool, and you can see that from the navigation. There will not be unnecessary links upon initial click, but will take you to its own dashboard for that specific event.
04. so long dreaded blue box
Currently upon entry to PX you are hit with the first image - a wall of helper text detailing what each section of the sidebar navigation does… Let's not.
I created the rectangle shape for the customer information section to replicate a physical Nando's Card, and all of that information is directly related to their Loyalty data. I've also gave the three main features their time to shine next to the card info.
There are a couple of bonus features; showcasing the customer's Lifetime Chillies next to the customer's name (big request from customers) and a Rewards list front and centre (another big ask from customers).
05. LESS LINKS, LESS CLICKS
Currently there are a host of links under the DBB when you look up a customer in PX. One of the main uses for users were seeing a customer's transaction history, so I've added underneath the customer info section.
I designed the table initially to replicate Paytronix's transaction history, and will work with users in testing to simplify this much further.
I've also added an initial filter as PX doesn't have one currently. This drives users insane as some customers want information about an order from 5+ years ago (not kidding).
06. Introducing Mr. modal
The current Adjust Wallet section is out of the viewport for the customer information, and users regrettably have to scroll up constantly to cross-check customer info.
I've kept the same vernacular with 'Adjust Wallet' as users are accustomed to this term, and fits their mental model.
The modal component works well with not being too disruptive to the user's experience, and they can still see sections of the customer information to cross check their work.
FEEDBACK SESSION ONE
The first feedback session went a lot better than I had hoped, and there was an air of excitement (which is funny because this is internal software). It started with the initial "WOW" of a much simpler interface followed by some extremely interrogating questions (another bad joke) by yours truly.
Although some of these statistics are fine, I am not ok with fine. This tool has to score near perfectly across the board otherwise I would be doing a disservice to my colleagues using it. Our findings were:
FIRST DESIGN EVOLUTION
USER FEEDBACK
I wanted to include the users as early as possible throughout the design process as this redesign is a legacy tool they use every single day. They have knowledge that I do not, and vice versa.
I went into the feedback sessions knowing the designs weren't completed yet, and although the feedback was initially good there was a lot of room for optimisation.
{ V1 }
{ V2 }
{ V2 }
{ V1 }
FINAL FEEDBACK SESSION
I hosted one more feedback session to gauge further user feedback for the newly designed solutions.
{ V1 }
{ V2 }
INTRODUCTION OF DESIGN SYSTEM
As stated in the hero section, this project was one of four pilot projects to help shape our newly formed design system.
I helped contribute by adding components such as cards, filtering, table drop downs, and pagination.
FINAL USABILITY TESTING
After running two separate feedback sessions with small groups, it was time to run 1 on 1 usability sessions with 14 users, again two from each team.
I tried to pick different users when I could, and the results were enough to make a grown man cry.
04 DELIVER
"Designing a product is designing a relationship." - Steve Rogers, ex Google
USER ACCEPTANCE TESTING
Final build testing results are in… Winner winner chicken dinner?
Paytronix - Crikey, 39 SUS
Loyalty Admin Tool - Excellent, 91 SUS
PRESENTING TO THE
CUSTOMER TEAM
I know it looks like there are only four people here but there were 70+ (I promise).
Telling the people my excellent "Luke's admin tool" joke, and walking them through the Double Diamond.
NEXT STEPS
A few features were not built for the MVP, such as filtering, transaction drop down, link to the Order's Hub, and impartial search. These are coming very soon.
05 MEASURE
"Design is not just about making things pretty, it's about making things work, then proving that they do."
- Yours truly, Sir. Luke Aris
IS IT IN USE NOW?
The Loyalty Admin Tool was finally turned on in March 2025, whilst we began to migrate over 10 million Loyalty accounts from Paytronix to Talon.One.
MEASURING PERFORMANCE
Whilst the tool was up and running, I still wanted to see if there were any improvements to make which would help our users even further conduct their daily jobs. I've been checking in with the teams weekly, via a Slack channel where users can ask queries and request new features. So far we have:
REFLECTIONS
This was my first end-to-end project at Nando's and I've learned so much whilst completing it. There are a few things I wish I did and have done in following projects such as:
- Including wider stakeholders much earlier in the design process
- Completing and sharing research findings sooner
- Definitely don't mute both your mic, and the meeting room's mic when conducting an interview
All in all, super proud of the work that I achieved collaboratively here and the users are also really happy using the tool. Time to celebrate, team.
THANK YOU.