CASE STUDY: HH
Objective
Improving 1-2 UX points on HH app: Case Study.
One week discovery crunch.

Pick one or two major problems and create some UX solutions, is the premise of this particular case study.
Of course, one must first define what a major problem is for a smartphone consumer app.
Research & Discovery
The whiteboard shows some of the research and micro studies into facts, figures and scenarios the user currently experiences. It was quickly evident a single approach to the experience was in place.
While no real data was available on the base metric of successful conversions, there were certainly many opportunities to improving flow, functionality and aesthetics (I've gone for the formative two points).
Dissection Study of Hungry House App: Entry + Selection pages compared across devices, an A to B page-flow, a mind-map and column containing public data on the company sourced from Desktop Research
The Approach
Knowing that lack-of-choice is predominantly the most valued complaint of the offer, the focus of UX improvement was to optimise the flow to the checkout.
The research had to be rapid, subjective and pose adequate options for later enquiry; the value in doing so sets the expectation that given additional time and resource gleans more accuracy and insight.
A quick 6-minute survey to my immediate network of participants catered for a high impact of immediate participation.
Guerrilla Testing
Building an online survey (Survey Gizmo) is fairly easy and distributable but  does provide you fairly quick results if you have willing participants. Which was not too hard in my case, since many friends and family do takeaway at some point.
Interviewing friends in their own attempts added to the board a number of ideas and thought-led patterns.
Surveying the people in your immediate circle and those happy to assist from your social channels; often provides beneficial results. The session is rapid, qualitative and memorable; you can quickly establish patterns to kickstart some initial ideas.
The questions designed were to expand on first impression experience, recall of particular moments and a big-picture-view of friction points along the imaginary user journey.
6 minutes to complete, anonymous and in tempo*; helps motivate participation. The objective was merely to setup markers of enquiry for later expansion or elimination and provide a design framework.
*Starting with easy quick checkboxes before moving onto more involved input the gradual warm-up is a persuasive technique.
Social-Public Data Research
In parallel, public information was explored to build a picture of the HH ecosystem. Press, YouTube, discussion boards, social media, statistical data and the web-apps source-code validation - brings a wealth of understanding; at the very least a rounded perspective of the service applications influence and its current standing on efficiencies.

Audible or Sound-only Approach
The information architecture (categories, groupings of data) was compared to natural voice transactions a customer might go through had the app be an actual person. By that we are looking at the taxonomy and category labels to affirm that they match the thinking model. Its a creative way of formulating folksonomy with the segmentation of content.
I created an audio path*, to establish the number of patterned steps to arrive at a conclusive food order. Usually, the scenario you may have experienced when taking and giving a food/drinks order. With the exception that you may also be ordering for others; design decisions can be formed from this basis and help establish an optimal funnel to conversion.

Rapid Prototype
Taking into account the aforementioned audio path, we next attend the visual information of on-boarding new visitors. This is best done by sketching, or paper-prototyping your intentions. https://goo.gl/euw5mR (Popapp)

Just one idea of grouping interactions in the lower ⅔ segment of the viewport on entry while ranking sign-in over registration as the first assumed interaction.
Once registration is complete the user gets straight back to the objective path of choosing. The use of imagery motivating the visual senses
The Mind-field
Walking through the journey; annotations suggest possible ideas like best in category sample dish - competition for such placements with the restaurants or indeed paid for.
Additional info below the main image indicate number of options available and either time to delivery or last orders countdown.
Following a registration complete, the "new" indicator in the menu identifies an update. It also identifies a location yet to be discovered - this would have shown the user that from their registration data this has made a 10% profile build; encouraging them to fill out more. They would have also discovered that they can preset their filters/sorting preferences.
Obviously, following the digital wireframe you can create your sitemap and better optimise the groupings. The isometric view will give you a better idea whether your branches are appropriately located and sized. It also provisions card sorting research should your categories need reviewing.
Possibly 50 or more cards to sort open/close and not having the 12 to 15 participants makes this approach a better candidate for later study. 
For the purpose of this short study an interactive wireframe was made the next step; assuming the sketching went through enough rapid iterations this is helps solidify the solution into an tangible format. https://goo.gl/r7tAHI (Balsamiq)
Interaction Design
Either the Popapp.in prototype or the Balsamiq prototype refined, would then lead onto more focused interactions, with that an Axure, InvisionApp, or indeed HTML5 render to follow. The definition of the micro-moments or micro-interactions are little opportunities for the UI to delight the users immersive engagement. 
Gamification of components, their relation to neighbouring atoms of elements should pleasure the senses of feel for a non-tactile interface. Here we utilise animation, scope, space, weight and size; assisting the framework of a UI designer to start ideating on a now known and founded structure.
In parallel we test functionality with the language of the design; the same way we have approached the research. Below is the part from the whiteboard that sat out to test the design against the objectives of the primary CTAs. Ideally both should match in importance, those that don't becomes noise or distraction.

When done right a balance is struck between Visual UI and UXD
An inconsistent entry and one that prevents further exploration is a poor implementation. Here the App forces the postcode as a mandatory requirement before proceeding to browse.
CTAs and Decorative elements of the same size; provides no sense of priority (website)
The lack of obvious CTAs blurs the available info
Personalisation Experience
As with any interface the basic categories of users are: first, frequent, loyal each representing different positions on the learning curve with their own intrinsic requirements and capabilities.
More features can be introduced to the UI over a period of usage.
Delivery Pipeline
Time planning, largely dependent on a number of static conditions; iOS + Android, one or the other and web-app + desktop.
Below is an Agile two week sprint of what the delivery pipeline may look like. The variation is in whether there is a pre-existing technology structure in place or if the approach is an iterative release, in which case perhaps a Kanban method is the better suit.
Let take an optimistic 4month-5month projection on a feature release new app. That's not being liberal with the Research, but a sketch to interaction + design and first cut.
Typical App Release plan for Interaction + Design for iOS and Android
Some simple points to remember, is that a staggered approach mean less chance of parallel errors. UXD starts first, one sprint ahead of FED; objective is the proof the concept and functionality.
Once FED start development (iOS), UXD builds on top of the wireframe spec the interactions.
At this stage any findings the iOS team find (Design issues/Errors) they can pass onto the Android team so as not to repeat the problem.
This also means that by the time Android development find their own issues, its already beyond the first set of issues and is a maturer fix.
One of the reasons iOS is factored as first to develop code is to account for the 14 day App Store submission, whereas Google Play Store has no human testing and can be published in around 4 hours.
Back-end development can start at the same time as UX since their work is largely decoupled from Front-End.
Similarly, Desktop is delivered in the same manner with variations in testing goals.
Some third-party services can assume the bench-testing (Xamarin). Ethnio.io is another Research service worth exploring if participation studies are short on time.
This is not to say that during anytime in the WIP stage some complex issues arise and often for unforeseen reasons - being able to adapt or knowing when to drop an original idea to save cost at development time or indeed design UI creation can be critical.
Retrospect & Conclusion
I think most risk can be mitigated by including your customer throughout the product lifecycle or at the very least a broad range* of internal candidates is better than no research test at all (NDA situ).
*Member of each department i.e BA, Dev, QA, UX, PM
The failing fast is about getting that real feedback to pivot direction quicker and the value is built-in since the engagement of users provisions its demand for wanted features. This can only be done incrementally.
For HH, getting more restaurants on their books would provide the much needed choice that their customers want in their local area. Perhaps diversifying outside takeaways and supporting local restaurants in general.
My suggestion here is that their target consumer think about dishes more than the single restaurant their meal is sourced from. Especially if to accommodate variety of taste; it would be more valuable to add dishes to the basket than have to restart the process per cuisine.
The UX features I've designed in plays on the idea of getting conclusion to their decisions faster and to the checkout. Reducing the need to go back on forth on content while using imagery to inspire informative decisions. 
Need a lift?