Through this presentation, I will try to convey my interface design process. I will show you how I think, how I work, what I pay attention to, and what I leave aside. So, grab a coffee and jump in.
No project exists in the infinite vacuum of imagination and incredible possibilities. When developing any project, we should always take into account many factors, such as:
Even when working on a design exercise, we should always be able to set the correct frames. Imagine that the project is 100%, the clients are real. And we have to understand that the product must enter the market. Also, remember that the product should have an MVP and, because of Google and personally myself, it must be innovative.
Since I'm doing a project within Google, for the beginning let's studying what Google is investing in now. What directions of technology development chose. What projects is more important, what less.
To do that, I've reviewed the latest Google I/O conference (the last one is the best), read the articles, asked friends who is working in Google and read the 44 pages of the CB Insights report: "Google Strategy Teardown: Google Is Turning Itself Into An AI Company As It Seeks To Win New Markets Like Cloud And Transportation", read Google mission: "Our mission is to organize the world’s information and make it universally accessible and useful."
From this study I conclude: Google relies on artificial intelligence, clouds, self-driving vehicles, and medicine. I use this data and focus on artificial intelligence. This topic is very close to me. At this moment, I constantly study this topic: take courses about AI, read books and articles (my favourite blog is waitbutwhy.com. What is yours?).
I chose the second of the three options: "Community issue reporting". I chose it, because I understand this problem better than the other two. I don't know much about school children or mentors. But I have a lot of urbanists friends who live in Moscow and make reports for the Moscow City Hall all day long.
Now we understand the strategy and business model of Google. Next step is to understand the main goal of our task. The brief says: "Design a system for a community of your choice, so members can report issues and track their resolutions". Following this brief, I can't understand either the company's strategy or the business model. Which means I will design layouts for the sake of layouts or interactions, which doesn't make sense in the real world. So, let's use "Job To Be Done" mythology to resolve this and ask a question: Why should any company hire Google for this job? And who can hire us (Google) and pay more for it?
In order to answer this question, it is necessary to understand what problems we are trying to solve.
We help residents report the problem within their neighborhood. So, it means we collect a large amount of data:
What we can do with this data and what benefits it brings to business (Remembering Google’s technical capabilities and strategy):
Total: optimizing the processes, reducing costs and increasing loyalty among people
I conducted a series of interviews with my urbanists friends, office-manager colleagues, a friend who worked in the Hackney administration, several friends who live in university dormitories and mom, who works as a teacher at school and often makes inquiries about mending furniture in her classroom. Everyone has about the same process.
I came to the conclusion that such a service can scale from school to a whole state. But for this task, let's focus on a small condominium in London, but will keep in mind that our solution has to be scalable.
We will sell our service to UK Government, specificity to Hackney Town Hall, London
So, we decided on who to sell. There is no strong business model yet. Now I need to understand why we create this service at all, what our goals are. What problem we are trying to solve?
Increase the rating of the Hackney area and loyalty to the local administration.
Plans for the future 5-10 years
Of course, this is all just an interaction design exercises, but it is necessary to think about this type of business processes and think about potential future of the service.
Make this service as scalable as possible (from office-managers to city-hall) and integrate it into the G Suit ecosystem.
Make an ecosystem of applications and services for Governments and public services like G Suit and call it "G Suit: Government". Thus, Google will be able to join in the system of contracts and tenders as Governments IT systems provider, capture the market and increase the profits.
Since we are trying to make the service for everyone, we can't influence the way the current system of work the condominium use at this moment. We just give them insights and metrics that show where there are shortcomings and how best to fix it.
Application addiction level
In the current era where people are actively monitoring their addiction level of technologies, it is very important to understand the level of addiction which I'm gonna design. Google itself provides additional features and tools that help people not to have an addiction to technology.
I think the application should be more native, than fun. It should not cause special addiction and desire to check the phone one more time. It is unlikely that the application will be used more than organize times a month.
Communication tone should be neutral but smooth, like: "don't worry" or "we will finish it soon" or "be careful".
We have to support the people, so our tonal ranges might appear like:
I believe that visual design should be in Google’s Material Design ecosystem.
The list of instruments and methodologies that I used when performing the task:
Main functionality (community member side)
I think that for such a task the most important is not the community member side, but the admin side. Admin can: view statistics, do reports, assign tasks to handymans, create new requests, ideas etc...
Unfortunately, I did not have time for such a complex and massive interface. Nevertheless, I tried to accomplish the main task: design a system for a community of your choice, so members can report issues and track their resolutions. But I also did a prototype of the admin part.
A prototype will be needed to complete the happy path.
To understand how the service will work, let's will take the story of fresh graffiti on the wall. Imagine that you are a person who lives in Hackney and saw fresh graffiti on the wall near your house. When designing the interface using Agile methodology, everything can be quickly changed. So, I will use this path as a guiding star, not a hard rule.
In our case, marketing has 3 main tasks:
Marketing strategy: see it, download it, report it, sorted.
On the front of the panel we will stick a film imitating broken glass and an inscription: Looks like this screen is broken — report it. Below is a VERY RAW prototype, just to explain the idea.
To begin, go to a friend who worked in public service in London and do CJM and co-ideation with him. Then come back home and do another co-ideation with my wife and friend, to get more ideas.
So, we have 3 types of users in our service:
Below you will see the personas. Collective characters that represent each of the three types of users. Of course, I can go deep and make sub-type for each type of users. Then make research on what sub-type of users are encountered more often in the city and build user scenario and interface based on these analyses. But it can take a lot of time and resources. Therefore, I created these personas, based on the interviews and my previous experience with interactions with such people.
As Alan Copper (creator of "Personas" methodology) once said:
"So we try to build software for everybody that does everything, and we ask ourselves what might anybody want and give that to them and that’s just the path to shit. Yes, I want to make everybody happy, but to make everybody happy, you need to begin by making just one person happy. And that was a fundamental insight."
How it works now
The handyman must first come and determine what the problem is. Then understand what tools he/she needs, what needs to buy. Then leaves, buys, comes again, repairs. If something does not fit, goes to the store again and can happen over and over. In addition, finding the right color of paint or the necessary wires can be very difficult, handyman losing a lot of time.
Layouts and happy path
Let's return to our example with fresh graffiti on the wall. Our community member John saw the fresh graffiti and decided to use a new service and report it.
Member side in Invision prototype is here
So, John sees the banner, downloads the app, logs in and scans the QR code.
To determine or switch to another area, you must enter the ID number (or scan QR code) of the area. It can be found on the district’s website or on advertising stands at the neighborhood.
Rate your area
We can also evaluate the opinion of community members about the area and public service before and after using the application. It will provide as with qualitative data. Therefore, after the first login, we provide John with a small questionnaire about the current state of the area. And in a 3 monthes, we will give this questionnaire to him again. Then we will compare the results and see how the new service influenced the general opinion of people.
I need to conduct additional research or find help from some Ph.D. Sociologists from ivy league universities in order to understand what questions will help us determine the emotional state of citizens in relation to their area. Unfortunately, I did not have time to do it, so I put approximate questions.
John makes the first entry into the application and sees a list of all the requests made by the residents of his district. He sees the title, subtitle, who created and the picture (optional).
All requests are divided into:
John can subscribe/unsubscribe to any request to receive notifications about how and when this task will be done.
To see only his requests, he needs to sort them through the drop-down menu. By default the list sorted by "Expectation date".
I also decided to add to this presentation ideas that did not pass for reasons of A/B tests or ideas that are not suitable due to the overall strategy of the application.
Do not divide into tabs
Initially, I brought all in one list, so that it was not necessary to switch btw tabs. But after a series of A/B testings, it became clear that it was easier for people with tabs.
They clearly see which ones are "Pending", which are "In progress", without additional sorting. Few people even click on the 3 points in the corner.
There was an idea to add an extra step: "Processing". It was needed to quickly go out of the "Pending". The manager could quickly understand the task and assign a handyman to it and the task becomes "Processing" (clearly not "Rejected"). Then, when the handyman had the time, he would go to perform the task and it would turn into "In progress". But the A/B tests showed that people do not understand why there are different statuses: some in tabs, others inside the request and why they are partly the same, and partly not. In general, everything became confusing.
The cards showed themselves very well on the A/B tests, but they take up too much space.
Give residents to vote for each requester, thereby creating an executive stack
The reason why abandoned the idea: It may happen that one house will be more active in voting than another and it will be serviced more often than the other, which will cause unpleasant consequences. In addition, voting will force people to additionally open the application and cause an additional habit.
Create a request
So, John taps on the plus button, and chooses from three options:
Creating ideas and requests is almost the same. The only difference is that the idea can't be private.
John choose "Maintenance" to report the graffiti. He sets title, location, explains the situation, makes a couple of pictures and sends the request.
Fighting with duplicates
By text and picture, we can identify duplicates and give community members the opportunity not to create duplicates.
We can use "Singular value decomposition" https://en.wikipedia.org/wiki/Singular_value_decomposition
Ideas request without private settings. Learn more about ideas below.
Categorization the request
There was an idea to make a categorization of what you want to report (Graffiti, elevator, entrance hall, door etc...). But people most often don't know exactly where to look for their request category (will the "fence be painted" in a "playground" or on a "street" category?). Categories are very important for general reporting. Categorization will occur on the admin side and will be automatically substituted based on text and pictures. If the categories are wrong, then the manager will be able to replace them.
But it is very important to make the requests as simple and fast as possible: see the problem - open the app - take a picture, describe the problem, set a location - send. That's it, THAT'S SIMPLE.
Request via chat
I wanted the requests to be made directly via chat with support. Perhaps requests could receive a chat-bot and independently assign it to the most appropriate handyman. But in general, this doesn't make much sense, because then it will be difficult to track requests, it will not be possible to see requests from all community members. In moments when the bot can't understand the request, it will send it to support, and since this is a chat, people are waiting for a quick response. That might load the support and will increase costs.
At the sight of another graffiti on a wall, a broken door or an unvarnished fence, many people experience discomfort and even anger. Therefore, there was an idea to somehow let people relax or laugh while making a request. Therefore, we wanted to give people to add text or make stickers on photos (as in Instagram stories). People love it and often do it themselves now, take a look at those two examples.
Of course 2 weeks to design such a system alone is almost impossible. Therefore, I will show only the main screens that show how to perform our happy path. Of course, there is a lot of work on the admin panel. Now it's just a raw prototype.
Manager task board (RAW)
Our manager Luis sees the request at the "Requests" section in the admin panel.
Google AI helps determine what the problem is. Categorizes the request, indicates which tools and materials can quickly help to solve the problem. It can also indicate the color of the paint, indicate when the last time this request occurred and who last performed it. For each tool, you can make a query, where to buy it in the nearest store.
We can use "Association rule learning", in particular, I suggest: Apriori, Euclat, FP-growth
To determine the objects and colors on the picture, we can use:
So, our manager Luis sees a request, leave a comment, change status from "Pending" to "In progress" and assign it to the handyman James (it is possible to assign multiple handymans or side partners).
John got a notification about the switched status.
Handyman side (RAW)
Handyman James sees the task and all additional comments and materials provided by the Google Lens and his manager.
Google Lens will tell him exactly what paint color and where to buy it, what tools to use, wall size, how long it will take to paint it etc... With the help of Google Lens handyman can speed up the work.
We must remember that our handymans are not very friendly with technology. Therefore, I've tried to make the interface as simple and understandable as possible. Remove all unnecessary.
James discovered that he doesn't have the right color of paint. He opens the app with Google Lens, point it to the wall to determine the color and the nearest store where to buy it.
James went to the store and buy it. He went to the shop just ONES. Come back, started work and changed status to "In progress".
John and Luis got a notification that his request is in progress status.
Also, James can leave a comment for residents. Can leave a comment for a particular house or apartment and notify them about anything.
James started to paint the wall. When he finished, he would switch the status to "Completed".
John and Luis get notifications about it.
John can open an app at any time to see his request. The Request page has the following structure:
When creating a request, community members could leave tags on where to look for the problem using AR. Handyman would find it faster. Everything is a bit like Pokemon Go and new Google maps. The idea is interesting, but not worth the money to develop and support.
I had a funny idea. People love to watch on how others are working. It is often not just to watch, but also to suggest exactly how to do it. Therefore, there was an idea to make a live-streaming of work, with likes and comments of course. Plus, for marketing purposes, stream it on Facebook and collect likes and comments there and increase the number of the app downloads.
But of course, this is a very strange thing for this type of application.
So, John received a notification that his request was fulfilled. The next day, when he returns from work, he will see that everything is really done (or will use the app to see it). He will open the application and leave an honest review.
Luis will get a notification that John left a review and happy path is done.
The review is very important for collecting data, so I highlight this function as brightly as possible and throw off additional notifications in the application and on the email.
Rating based on 3 data:
If community members were unhappy with the work and remained dissatisfied while writing a review, Google understands this and suggests creating another request, based on it.
Separate "Review" page
There was an idea to make a separate page for the review, and not to make it down below and hoped that people would scroll to them. But this would be an additional tap and in most of the Google applications reviews are located directly on the product or service page, so I decided to go with Google’s standards.
The review is an important part of the whole process. Therefore, I wanted to somehow promote and motivate community members to leave feedback. But the idea seemed a bit intrusive, it must be carefully tested (especially where this promo should be, how much attention it should attract etc...), I have no time left for this test.
A place where community members leave their ideas about improving the area. People will feel that their ideas are being listened to and executed, which will greatly increase loyalty to the government.
Every idea has a deadline: no more than 60 days. If after this time the manager did not process the idea, then it automatically disappears and becomes unaccepted. The creator receives an email notification about it.
If the idea has been accepted, then it automatically falls into the maintains section to the pending status.
Community members can vote for each idea. Numbers of votes built as a rating of posts on Reddit.
Rejected ideas on Ideas screen
Raising money for the idea
People often asked this feature in ideas. It seems to me that it is correct. But in general, it will be too much for such a simple application. We can build a whole Kickstarter from this, which is not a good idea. And the goal of this section is to show that the local administration is listening the people. Thus, we are moving away from the main goal, so it makes no sense even to design the prototypes.
Voting for an idea through
thumb up/thumb down icons
The idea proved to be quite optimistic on A/B testing. But in general, the concept of thumb up and thumb down should not create a general rating. With such a vote it would be good to show how many community members voted for it, and how many against (like on youtube). And I wanted to make an average rating because I consider it more correct for this purpose.
Therefore, I leave the arrows.
Handyman or manager can send notifications to community members if they need to pay attention to something. For example, when the house is painted. Or they can warn you of more serious danger, for example, if someone a that they have water leaking or someone smells gas.
Email and app notifications:
After a series of interviews, I learned that all public services have an understanding of an emergency request. Most often there is a list for which he is considered an emergency. This list can be managed in the admin panel.
Let's take our Hackney area and see the list and add it to the application for emergency situations:
These requests are immediately falling Into the "In progress" section. Usually, managers and handymans respond to this requests within 10 minutes. In the application, we immediately offer to call and tell about the situation. We provide instructions on what to do to protect yourself.
In case of an emergency request, we also ask to leave contact details. This will be especially useful if the problem is in the apartment.
This is just the beginning
In addition to standard metrics (speed of performance of tasks, quality of work, which categories are the most problematic, what people pay more attention to, the level of satisfaction of residents etc...) that we can let go in this task, I would like to show the most interesting ideas.
Let's take the same example of our graffiti. It appears in the same place all the time. Handymans even already have the same paint for this place. This happens everywhere.
So, let's use big data, machine learning, and predictive analytics to indicate that graffiti will soon appear here again accurate to the day. We can send this info to the police at this case.
See real story, happened in Detroit, MI at below gif
But just take another case. For example, elevator constantly break down (it is a real situation in my house ). Using all the same tools, we can predict how much time will pass and it will break again. And in advance to send handyman to fix it. community members will see that the elevator is working and never break again. MAGIC.
In some cases, we will win time, and in some life (for example, if you use this service in a scale of the country and prevent the bridges from falling).
Google already has such technology to do that.
Just as part of the ecosystem can be offered to make smart storage. In the storage where paints, tools and so on are stored ... Add cameras and see if there are any necessary tools through a mobile application. Also, if bought paint, then noted in the application. If used paint, then noted in the application. So you can control the amount of paint, the number of instruments and so on... And automatically/manually buy what ends.
Thus, we can help the handyman not to go to the store at all and immediately go to the store and vice versa.
Just like smart refrigerators are doing now it now
Sharing data between neighborhoods
We can analyze the knowledge and skills of district staff. As well as materials and tools that were used in each of the areas. If we share these categories between all areas of the city, then we can transfer materials and knowledge of people between areas.
For example, an electrician in one district is well versed in transformer booths. If someone needs help, from another neighborhoods, they can call him, and not pay to outside partners. Or send some materials, such as paint or tools etc... It will reduce costs, since everyone works for the state.
Since tasks and requests are created inside admin panel, which itself categorizes all requests and understands the problems, we can transfer data about all the problems that exist in the city. May make reports for the government on the condition of houses, districts, condominiums etc...
Other rejected ideas
The idea is to make gamification to motivate community members to create requests. The idea is good, but not for this application. The whole idea of gamification is based on the Hooked model (Nir Eyal): Trigger - Action - Variable reward - Investment.
If you follow this concept, you do not need to do gamification at all:
But add gamification here and people will be madly chasing new badges and statuses, which will only lead to excessive addiction.
Community members rating
Goes as part of the idea of gamification. The more often you make requests or leave and ideas, the more points you get and the higher your rating among neighbors. The idea is strange, but there is a place to be mentioned. Again, it can be addictive and unnecessary rivalry.
I had a funny idea to make an application icon in the form of Mario.
I want to say that the task was interesting. In my opinion, all layouts still look a bit raw. Much more work is needed and tested. Need more people and time to check the viability of the idea. Probably, I would also need to design screens with statistical data for the managers at admin-side. But to design it, will require a large number of days. And I had 2 choices: 1. Either to design nice-looking, but absolutely meaningless graphics from my head, without analyzing, without understanding how to show them correctly, which visualization is correct. 2. Focus on a global goal: create a system for community members to report problems. I took the second option
The more I did this task, the more I realized that the prospects for such an application are not as much as we would like it should be. Especially from the budget and benefits for Google itself. And I would recheck the idea, maybe change the strategy, test it, find the best way to monetize and benefit for Google more. And only after that would send the service to the development. At this stage, I would say "NO" to such a project.
The most useful discovery is to sell the premium version of Google Lens (perhaps Google is already doing this). Specifically tuned to handymans so they can buy certain tools faster and more conveniently.