← Work/2023 – 24 · Product Design, User Research, User Interviews

Basket

A grocery delivery app in Amman, redesigned from onboarding to checkout so that every screen can be read and every step is easier to take.

1 h
Delivery window, in Amman
8
People on the team: product, development and marketing
7
Flows redesigned, from onboarding to refunds, plus the home screen
20
Recent sign-ups interviewed about sign-up and onboarding

Impact figures are those reported in the original case study (updated 10 June 2024). It does not record how each was measured, so treat them as reported, not audited.

In 60 seconds

Role
Product Design, User Research, User Interviews
Timeline
2023 – 24
Team
8 people: 2 in stakeholders and product, 4 in development, 2 in marketing

The challenge

The old app worked but could not be scanned: one green scheme, text laid over product photography, and calls to action that did not stand out.

The insight

Interviews and tests kept returning to the same points: people could not find the primary action, could not read text over images, and wanted guidance through onboarding. The checkout heatmaps showed clicks away from the main action and drop-off at payment selection.

What I did

Ran interviews, usability tests, a four-app benchmark, heatmaps and an A/B test, then redesigned seven flows and the home screen, from onboarding to refunds, around a clearer hierarchy, readable text and shorter screens.

20

Recent sign-ups interviewed. They asked for more guidance in onboarding and for calls to action that stand out

Interviews · Qualitative

15

New users did sign-up and onboarding with screen recording and eye tracking. Users struggled to find the primary action in cluttered layouts

Usability test · Mixed

3

Things the checkout heatmaps showed: clicks away from the main action, drop-off at payment selection, almost no use of the promo code field

Analytics · Quantitative

+40%

Checkout completion, from the original case study, which names an A/B test of the old checkout against the new as its method. The sample and run period are not recorded

A/B test · Quantitative

The platformWhat Basket is

Groceries from the stores you already use, at the door in as little as an hour.

Basket delivers in Amman, mostly in the west of the city, with plans to extend the area. My part was product design, user research and user interviews, with a team of eight: two in stakeholders and product, four in development and two in marketing.

01

Select a store

Local and national grocery stores to choose from.

02

Add items

Browse the store's inventory and fill the basket.

03

Place the order

Finalise the selection and order.

04

Fulfilment

The order is dispatched through an Uber-like system to shoppers, who receive it on their phones and fulfil it.

The problemWhat was wrong

It worked, and nobody could scan it.

Six problems with the old app (version 1), each of which came up again in the research.

01

Look and feel

One green scheme, with text over product photography. Functional, but every screen looked alike.

02

Text and messaging

Text on top of images was hard to read, and the key benefits (savings, service variety, availability) were not prominent.

03

Onboarding

A consistent pattern with no strong visual cues to guide anyone through it.

04

Call-to-action buttons

Sign In, Register and Continue as guest sat in the right places but did not stand out against the background.

05

Imagery and icons

Product images on one green background: informative, and monotonous.

06

Density and clarity

Several pieces of information on each screen, which felt dense and slightly cluttered.

The researchWhat did it reveal?

People could not find the main action, read the text, or tell what to do next.

Five findings from interviews, usability tests, a benchmark and the checkout's own heatmaps. Each one is what we saw and the evidence behind it; the design response comes after. Sample sizes are as written in the original case study.

01

People were confused by the lack of hierarchy and by calls to action that did not stand out.

Interviews· QualitativeAsked for more prominent, distinguishable calls to action.
Usability test· MixedUsers struggled to find the primary action in cluttered layouts.
02

Text laid over product images was hard to read, so important messages were missed.

Benchmark· QualitativeSuccessful apps separate text from imagery.
Interviews· QualitativeUsers preferred clearer text placement and readability.
03

Onboarding lacked visual cues and interactive elements to carry users through the steps.

Usability test· MixedUsers were more satisfied when onboarding gave visual guidance.
Interviews· QualitativeUsers said they wanted more guidance.
04

One green scheme and heavy product imagery made the experience monotonous.

Benchmark· QualitativeApps with livelier colour and varied visuals saw higher engagement.
Interviews· QualitativeUsers favoured more visually diverse designs.
05

Screens with many pieces of information felt dense and raised the effort of reading them.

Usability test· MixedSimplified layouts led to quicker task completion.
Survey· QuantitativeUsers reported a preference for cleaner, less cluttered interfaces.

How we knew

Interviews

Qualitative

20

Users who had recently signed up, about the sign-up and onboarding. Fifteen regular delivery users were interviewed about addresses, and twenty more watched while they browsed and added items (contextual inquiry).

Usability test

Mixed

15

New users completing sign-up and onboarding with screen recording and eye tracking. Ten frequent users did the same for address tasks.

Benchmark

Qualitative

4

Instacart, Delivero, Talabat and Uber Eats, read for onboarding, calls to action, address management and checkout.

Analytics

Quantitative

Heatmaps on the checkout pages with UXCam: clicks on areas unrelated to the main action, drop-off at payment selection, and almost no use of the promo code field.

A/B test

Quantitative

Checkout version A, the current design, against version B, with a clearer hierarchy and prominent calls to action. Version B had the higher completion rate and the lower drop-off.

Design judgement

Qualitative

A journey map of the checkout, and card sorting for how people group and look for items.

The responseHow the design responded

Five findings, and what each one changed.

Any figure on a card is labelled as reported in the original case study. It does not record how each was measured.

01

A clearer visual hierarchy

What we saw
People were confused by the lack of hierarchy and by calls to action that did not stand out.
Evidence
Interviews· QualitativeAsked for more prominent, distinguishable calls to action.
Usability test· MixedUsers struggled to find the primary action in cluttered layouts.
So we
Rebuild the hierarchy so the important elements stand out.
What happened
Reported: +28% click-through on calls to action.

02

Text off the photographs

What we saw
Text laid over product images was hard to read, so important messages were missed.
Evidence
Benchmark· QualitativeSuccessful apps separate text from imagery.
Interviews· QualitativeUsers preferred clearer text placement and readability.
So we
Dedicated areas with contrasting backgrounds for text; key benefits in larger type with icons.
What happened
Reported: +20% readability, and +35% understanding of the benefits.

03

Guide people through onboarding

What we saw
Onboarding lacked visual cues and interactive elements to carry users through the steps.
Evidence
Usability test· MixedUsers were more satisfied when onboarding gave visual guidance.
Interviews· QualitativeUsers said they wanted more guidance.
So we
Progress indicators and tooltips, and a distinct visual identity for each onboarding screen.
What happened
Reported: +35% to +40% onboarding completion, depending on the measure.

04

A more engaging look

What we saw
One green scheme and heavy product imagery made the experience monotonous.
Evidence
Benchmark· QualitativeApps with livelier colour and varied visuals saw higher engagement.
Interviews· QualitativeUsers favoured more visually diverse designs.
So we
A more varied palette, illustrations and icons for categories and actions, and varied backgrounds to tell sections apart.
What happened
Reported: +30% visual engagement, and +35% on feedback about visual appeal.

05

Less on each screen

What we saw
Screens with many pieces of information felt dense and raised the effort of reading them.
Evidence
Usability test· MixedSimplified layouts led to quicker task completion.
Survey· QuantitativeUsers reported a preference for cleaner, less cluttered interfaces.
So we
Ample white space, with the most important information first.
What happened
Reported: -20% cognitive load, and -14% drop-off.

01 of 07

Sign up and onboarding

Version 1 opened on an icon grid and an illustrated carousel. Version 2 opens on full-screen food photography, with Sign In and Register in reach and a phone-number entry that has its own screen.

BeforeBefore: version 1, the old app · 6 screens, each on its own
AfterAfter: version 2, the latest app · 5 screens, each on its own

What improved, part by part

The first screen

A tiled grid of grocery icons with the line Shopping is our thing and a Continue as button.

Full-screen food photography with the basket logo, and Sign In, Register and Or continue as guest at the foot.

Explaining the app

A carousel of illustrated screens: groceries in 10 minutes, the largest supermarket brands, your time is valuable, and coupons.

Straight to the choice to sign in or register; the promise is left to the photograph.

Phone number

No number-entry screen among the version 1 screens.

An Enter your number screen: a country-code field, a green Continue, and the numeric keypad open.

Other ways in

Sign In and Register buttons under the carousel.

Google, Facebook and Apple below the number field.

When something is wrong

No error screens among the version 1 screens.

Designed errors for a wrong code, an email problem and a user that already exists.

Insight

Version 2 puts the way in first: the number field and the main button are the screen, instead of a carousel to get through before them.

Reported in the original case study, not audited

  • +35%Onboarding completion
  • +40%Satisfaction with onboarding
  • +22%Sign-up rate

02 of 07

Addresses and locations

Version 1 asked for an address on a pin and a description. Version 2 starts from a full map and search, then asks for the details a courier needs, and keeps the result in a saved list.

BeforeBefore: version 1, the old app · 9 screens, each on its own
AfterAfter: version 2, the latest app · 5 screens, each on its own

What improved, part by part

Finding the place

An Address delivery screen with a search field over a mostly empty map and a pin to confirm.

A full map with Search for an address, and suggestions as you type.

Current location

Not offered on the address screens.

A banner that says Current location, location service is off, with a Turn on button, and a locate button on the map.

Details for the courier

A single Description box and a Save this location for next orders switch.

An Address info form: address type (Home), building name or number, entrance or staircase, floor and apartment, and How do we get in.

Confirming

A Confirm pin location button.

A Save and continue button under the form, with the delivery address shown on a map above it.

Saved addresses

Saved with a switch, with no list screen among the version 1 screens.

A Saved addresses screen, with an empty state that explains why to save one and an Add New Address button.

Insight

Version 2 splits the job in two: choose the place on the map, then describe how to reach it, so neither screen asks for both at once.

Reported in the original case study, not audited

  • +25%Engagement
  • +30%Satisfaction

03 of 07

Cart

Version 1 showed one store's cart in full. Version 2 adds a Your Carts screen that lists every store's cart, so a shopper with an Express order and a supermarket order can see both.

BeforeBefore: version 1, the old app · 4 screens, each on its own
AfterAfter: version 2, the latest app · 2 screens, each on its own

What improved, part by part

Many carts

A single store's cart at a time, headed by the store and its total.

Your Carts: a card for each store (Express, Carrefour, Burger Maro), with thumbnails of what is in it, +15 for the rest, and a bin to clear it.

Getting back to a cart

Through the store.

A green View cart button on each card, with the item count and the total.

Finding a cart

Not applicable with one cart.

Tabs along the top: All Departments, Express, Supermarkets and Food.

Nothing in it

No empty state among the version 1 screens.

A Cart is empty screen with a cart icon and a line of copy.

Insight

Version 1's cart stayed inside one store; version 2 lifts it out, so the baskets built across stores are visible together.

Reported in the original case study, not audited

  • +35%Cart additions, reported for the item listing redesign

04 of 07

Checkout

Version 1 laid every choice out on one long screen. Version 2 reduces checkout to a short list of rows, and each choice opens in a sheet of its own.

BeforeBefore: version 1, the old app · 9 screens, each on its own
AfterAfter: version 2, the latest app · 5 screens, each on its own

What improved, part by part

The screen

One long scroll: delivery location map, delivery time, instructions, payment, tip and the payment summary, all open at once.

A short Express Checkout: delivery address, delivery time, payment method, instructions and promo code as rows, with a Total and a Confirm & continue.

Delivery time

Day chips and slots inline, with a Confirm button.

A Schedule delivery time sheet: a row of days, a list of slots with the fee on each (JD 1.99), and Confirm time.

Payment

A list of methods inline: a saved card, New Credit Card and Cash.

A Payment method sheet: Bring card machine and Cash on delivery, then Pay with card with Edit, Delete and an Expired flag, Add Card and Confirm.

Promo code

An inline field with an Add button.

A Promo code sheet: the code, Redeem, 10% discount applied, and Done.

Free delivery

Not offered on the checkout screens.

An Unlimited free delivery card with a Try Prime button inside the checkout.

Insight

The heatmaps showed clicks away from the main action and a drop-off at payment selection. Version 2 shortens the screen that people have to read and moves each long choice out of it.

Reported in the original case study, not audited

  • +40%Checkout completion
  • +35%Satisfaction
  • -25%Drop-off
  • +28%Call-to-action clicks
  • +20%Promo code use

05 of 07

Order tracking

Version 2 turns tracking from a status page into a place to act: change the order, reach the shopper, change how you pay, or reschedule.

BeforeBefore: version 1, the old app · 10 screens, each on its own
AfterAfter: version 2, the latest app · 5 screens, each on its own

What improved, part by part

Status

A map and a five-step progress bar, with a title that changes: Order Review, Checkout, Checkout complete, Arrived at customer.

A map with the delivery pin, and a four-step bar: Placed, Shopping, In the way, Delivered.

Changing the order

A notice that you will be told of any changes, with a Review button.

A banner that there is still time to change the order before shopping starts, with Update items list and Cancel.

The shopper

A Chat with the shopper button along the foot of the screen.

A shopper card with the name, where they are, and message, call and WhatsApp icons.

Payment

Changed on a separate screen.

The payment method on the tracking screen, with a Change button.

Rescheduling

Not offered on the tracking screens.

A menu on the order with Reschedule.

From the home screen

A track-order widget in the wires.

An Order status card on the home screen, with the store, items, total and the progress along the way.

Insight

Version 1 told you what was happening to the order; version 2 lets you do something about it while there is still time.

Counted from the screens

  • 3

    ways to reach the shopper: message, call and WhatsApp

  • 3

    things to do to an order before shopping: update items, cancel and reschedule

  • 4

    stages in the progress bar

06 of 07

Complaints, refunds and cancelling

Version 1 showed the order being picked, item by item, with replacements to approve. Version 2 adds the other side: raising a problem with an order, taking action on what was unavailable, and cancelling with a reason.

BeforeBefore: version 1, the old app · 7 screens, each on its own
AfterAfter: version 2, the latest app · 5 screens, each on its own

What improved, part by part

Raising a problem

A My Items screen during shopping, with groups for replacement, refunds, shopped and adjustments.

An Order Complain form: each item with a quantity stepper for how many were affected.

What went wrong

Replacement choices and Other options on each item.

Select one or more product issues, in a sheet: Damaged Item, Missing Item, Expired Item, Wrong Item or Other.

Evidence

Not asked for.

An optional Attach Photo with an Upload button.

Getting it right

Approve on each replacement.

Select at least one reason is shown in red when the form is sent without one.

What happens next

A list of possible replacements with a search, and the option to refund an item.

A Take Action screen: item replaced, item canceled and refunded, and for each unavailable item a Refund or a Replace Item button.

Cancelling

A Refund this item button on an item.

A Cancel Order sheet that asks for a reason, from High Delivery Fees to Change in payment method, with Cancel and Confirm.

Insight

Version 2 gives the customer a way to say what went wrong and to choose the remedy, instead of only approving what the shopper proposes.

Counted from the screens

  • 5

    product issues to choose from

  • 7

    reasons for cancelling an order

  • 2

    actions on an unavailable item: refund or replace

07 of 07

Item listing

Version 1 listed one store's items in a grid with filters. Version 2 groups search by store and aisle, with a loading state and a way out when an item cannot be found.

BeforeBefore: version 1, the old app · 4 screens, each on its own
AfterAfter: version 2, the latest app · 5 screens, each on its own

What improved, part by part

Searching

A search inside one store, with category tabs and a grid of items three across.

A search across Stores and Restaurants, with sections by store: Express, then Carrefour, each with View All 10 Items.

The list

Prev and Next buttons, a filter sheet and a free-delivery banner.

Aisles with their own tabs and chips (Dairy & Spreads, Eggs, Butter and Ghee), with the headings Eggs and Butter, Margarine & Ghee.

Loading

No loading state among the version 1 screens.

A Search Shopping List loading screen made of skeleton tiles.

Quantity

A stepper, with a bin to remove.

A stepper on the item (2 ct), and a View cart pill with the count at the foot of the screen.

Offers

Red discount badges on items.

Offers with a count (363) and See All, with the old price struck through beside the new one.

When it is missing

Not covered.

A Not available label, and Can't find item? with Ask for product availability.

Insight

Version 2 organises by where things are, store and aisle, so a long list is read in sections instead of paged through.

Reported in the original case study, not audited

  • +35%Cart additions
  • +40%Satisfaction
  • +30%Engagement
  • +25%Navigation efficiency

The home screen

Home, part by part.

Version 1's home was dense: promotions and categories did not stand out, and the images were too small to engage with. Version 2 changes what the first screen offers: where to, search, the services as tabs, category shortcuts, an Order again row, restaurants, and the status of an order in progress.

Before: version 13 screens, each on its own
After: version 23 screens, each on its own

What improved, part by part

Where to

An ETA chip with the address (30 to 40 mins) over the greeting.

A Home chip with the address and a search for groceries and restaurants.

Services

Large tiles: Basket Express with its 10-minute clock, Basket Hyper, Supermarkets, Flowers, Locals and More.

Tabs for the services: For you, Express, Supermarkets, Locals and Pharma.

Categories

Not on the home screen.

A grid of category shortcuts: Offers, Local Items, Water, Produce, Bakery, Dairy & Spreads, Milk, Snacks, Ice-cream and Coffee & Tea.

Shortcuts

A Hello greeting by name.

A shopping-list banner (Try Shopping List Save Time) and an Order again row with a plus on each item.

Restaurants

Not on the home screen.

An All Restaurants list with cards: a photo, a Featured or discount tag, the Prime mark, distance, hours, fee and rating.

Offers and orders

Not on the home screen.

An Unlimited free delivery banner for Prime, and an Order status card with the store and progress.

Navigation

Home, Orders, Help and Account.

Shop, Restaurants, Orders and Account.

Counted from the screens

  • 6

    parts of the home screen changed

  • 6

    service tiles in place of a row of tabs and a grid of small icons

  • 4

    navigation tabs, down from five

Reported figuresWhat the reported numbers say

Where the reported lift is highest, and what version 2 added.

The charts are the Cortex design system's own bar and waffle components, running live. The lift figures are those reported in the original case study, not audited; the screen counts are version 2’s screens, grouped by stage.

Insight 01

Reported satisfaction rose wherever it was asked

Between +30% and +40% in each of the four flows that report it, highest in onboarding and the item listing.

Insight 02

The biggest reported jump is where a decision is made

Checkout completion, at a reported +40%, is the largest completion or engagement figure reported, ahead of onboarding at +35%. It is also the screen that version 2 shortened most: one long scroll became five rows and three sheets.

Insight 03

Version 2 grows the part after the order

Almost three in ten of version 2’s screens come after the order: tracking with its change, cancel and reschedule actions, complaints, refunds and cancelling with a reason. Version 1 showed the order being picked; version 2 lets the customer act on it.

HandoffIn Figma

Every flow drawn with its notes for development.

The Figma file keeps the old UI (imported from XD) beside the new Basket app, the components, the landing page and wire thoughts. Flows are drawn as connected boards with developer notes on the screen they belong to, such as the restaurant cards list with Add to favourites.

The Basket App - Consumer file: the Restaurant cards list and Add to favourite flow, with its developer notes.

Note 01

Feedback you can feel

Tapping the heart triggers haptic feedback, so a person knows the favourite really happened. Removing one does the same.

Note 02

Cards that adapt

The card header is an image slider when a restaurant has more than two photos, and a still image as the cover when the owner has only one.

Note 03

Two cover cases

Case A is a still image. Case B is a still image with the logo placed on it. Each is named on the board so development builds the right one.

The itemOne card, many places

An item is set differently in a restaurant, an express shop and a supermarket.

The design system's item components cover the placements. A restaurant item adds sizes and options with a running total on the Add to cart button; a supermarket item stays a quick plus and a quantity.

ImpactBefore, after and what was reported

Six things that changed, and the figure the case study gives for each.

Look and feel

One green scheme, text over product photography; every screen looked alike.

More vivid colour, varied backgrounds and larger type.

+30% engagement, +25% retention

Text and messaging

Text over images; key benefits not prominent.

Bolder text on solid or gradient backgrounds; key benefits stated plainly.

+20% readability, +15% conversion from messaging

Onboarding

One pattern repeated, with no strong cues.

A distinct identity for each screen, and graphics that keep interest.

+35% completion, +40% satisfaction

Calls to action

In the right places, but they did not stand out.

More contrast, easier to find and press.

+28% click-through, +22% sign-up rate

Imagery and icons

Product photographs on one green background.

Illustrations, icons and background patterns that support the text.

+32% visual appeal rating, +18% time on each screen

Information density

Several pieces of information on each screen.

More space, with short, clear messages.

+35% clarity, -14% drop-off

These are the figures the case study reports for the redesign. It does not document how they were measured, and its own conclusion says the changes are likely to give a better onboarding experience, so they are shown as reported rather than as verified results.

In shortWhere it landed

A redesign that put readability first, and made the path to checkout clearer.

The case study reports higher engagement (+30%) and retention (+25%), and credits the work of the stakeholders, product, development and marketing teams. Home and tracking were still being worked on.

© 2026 Alhasan HosniHome