Data-driven design at du Telecom.

Redefining the telecom's sales & services digital journeys, whilst fostering a culture of driving design through data

Background

du is a telecom company in the UAE with a subscriber base of 8 million across mobile and fixed lines (TV & internet).

In 2019 and 2020, the digital team worked grounds-up to design, develop and deliver the new du app & My Account (for self-service) and the new shop (for sales) for both personal use & enterprise customers. Post MVP, the efforts were forked into two stream- The major focused on delivering the rest of the roadmap, another smaller one focused on optimising what is delivered.

My role

As UX Design Manager from 2018 to 2020, apart from owning the design process for the sales and services offerings on web and app, I wore multiple other hats.

📢 Evangelise UX

Alongside the Creative Director, I worked as an UX evangelist in the business with the goal to elevate it from an exclusively tech capability to an org-wide capability.

🧠 Reposition the team from a wireframe factory to a problem-solving unit

By involving UX team upstream in the process and by consciously speaking the language of the business.

💪 Team building

Working along with the Creative Director and another UX Manager, we grew the team by 50% with clear focus areas for new hires. We realigned the whole team into smaller, tighter ‘squads‘ with UX, UI and FE Dev coupled together.

💻 UX ops

As a hands-on UX Manager, I would often tweak various parts of our team ops e.g. UX delivery process, design system governance process, design-dev handovers, etc.

📈 Continuous improvement

By bringing analytics and metrics within the design team’s way of working and inculcating a culture of experimentation.

MVP is out! What now?

Defining and designing for MVP and the following efforts in delivering the rest of the roadmap had its own set of challenges. As part of this case, I will focus on the experience optimisation efforts post-MVP.

A separate stream for optimisation was not initially planned. A variety of causes and efforts culminated into the setting up of the ‘Optimisation squad‘ in du’s digital team. One of it was also, me taking a business course called d.MBA around the same time. More than anything, I learned to have empathy for our business the same way we have it for our users.

Cross-functional optimisation squad

A lot of lobbying granted the squad an autonomy to decide what to build, how to build it, and how to work together while building it. Along with 50% of Product Owner’s time, the squad comprised of myself as UX Designer Manager, a Data Analyst, a UI Designer, a Front-end developer, a Back-end developer and a QA engineer.

Alarms blaring

There were multiple problems looming large- Dismal traffic to eShop, faltering conversion in the funnel, dipping TNPS, out-of-cadence with segment (business owners), team structure (split between Front end Dev and IT). While the new structure of the squad, bridged some existing gaps between teams, we jumped in to fix the high-visibility problems.

Let’s talk about one of them.

ID renewal, as a case

Showing you a schematic of the user-flow to simplify the discussion. On a high level the process of buying a du Plan has 4 steps- First you compare and select a product, then you review cart, then you put in your details, then you finish with your credit card details. All connections need to be legally bound to a ID (Emirates ID in UAE). Submitting the accurate ID details is a crucial and mandatory step in the journey.

One of the major problem was that a lot of orders couldn’t be fulfilled at customer’s doorsteps because of inaccuracy of the details entered during this step.

Capture the Emirates ID through OCR. Simple?

It’s never as straight-forward as that. OCR as a technical solution has been stuck with legalities. We had to find a tech design solution.

“Diagnose with data. Treat with design.”

— Chandra Narayanan, Co-founder of Sundial

Modus operandi

⚒️ Hunt & gather

As opposed to super-fast delivery, the squad strived to find "diamonds in the dirt" through analysis and speculation.

👩‍💻 Groom & execute

Investigate through open-ended design methods, engage in fast design and development based on a hypothesis.

📊 Monitor & iterate

Tracking of agreed metrics, analysis and subsequent action based on the outcome.

🔁 Repeat

Diagnosing with data

Presence of Data Analyst gave the squad very early visibility in the specifics of the problems. We had 2 very peculiar insights-

≈7% of rejections
were for UAE nationals

≈6.5% of rejections
were because of incorrect ID Expiry Date

These data points informed us about what is happening, but nothing about why. Or what to do about it?

Treating with design

Once we know the what is happening in detail, we can engage in creative, open-ended problem solving (aka design).

In the new ways of working in the optimisation squad, to ensure constant user involvement we had recurring sessions with users every 4 weeks. For every session, I would formulated the research brief, define participant demographics and the methodology, and over see a 3rd party research partner run the session. In these sessions, the research tools were Contextual inquiry & Diary studies.

While our engagement model was not great, the idea was to get as much user involvement as possible. There were 2 main insights from these sessions-

Not everyone uses their Emirates ID while buying a du Plan.

This insight though not relevant for new buyers, solved a major problem that I had encountered in my previous squad.

Users need the Emirates ID card only for few details.

The incorrect expiry date was because it was the only field that users don’t remember and was on the back side of the Emirates ID.

Hypothesise, test & iterate

While the user sessions were underway, our first iteration was to have Confirmation in a UI which visually resembles the Emirates ID. This did not effect the accuracy much.

Meanwhile the insight from user research informed that we need to club the fields that one doesn’t remember.

The 2nd problem - UAE nationality error was due to autofill. The same page had the delivery address form and the autofill assumed Nationality as one of the address field. It was simply handled by disabling the browser auto-fill for that field.

Impact & learnings

The optimisation squad ran 40 odd experiments in the span of 8 months, of which many were futile. Others were able to influence few business or design metrics in some way or the other. However, the biggest impact, was the change in the way the design team embraced the use of data. It was fulfilling to witness designers open up various dashboards and engage in a design critique session!

The data maturity in the design team was far from where it should be. We still struggled to run multiple experiments in a go. In the course of 8 months, we could comfortable rely on data to optimise existing product but could not inform any new work stream.

Design team

Adam Marston (UI Design), Åsa Andersson (Design Director), Chayan Deb (UX Manager), Huzefa Hussain (Front End Dev), Mas Matar (Studio Manager)