How to improve the product when it seems easier to make a new one instead
I am working part-time as a designer for a medical startup. It has already taken 150 hours which is equivalent to 1–2 months of full-time work. It has been going on like that for 1 year. During this period we have moved on slowly but steadily. Although we have launched only a small part of what was designed, some metrics have considerably improved:
- Time on site increased from 2 minutes to 20 minutes, i. e. 10 times.
- Page depth increased twice
It is clear that longer time on a site is not always a good metric, e.g. in case of search engines. As our service suggests programs from different medical insurance companies, people need time to compare their quality before buying 1 year medical service. In other words, it is a classical online market for insurance programs but with a longer buying cycle.
Here is the story of how we achieved this result.
Overview of problems
Users coming from search engines wanted to buy the program but they didn’t understand the advantages or didn’t trust the service and chose to buy directly from insurance companies or not to buy at all.
Here are some reasons why discovered by me and CEO:
- Problems with the content and design
- It was difficult to understand the advantages of insurance programs and especially of buying from MEDO.
- It was difficult to choose and compare programs.
- The program included many different characteristics but was designed as a simple text without any styles or visual elements.
- Colors were bright but accents were put in the wrong way.
- The flow of buying wasn’t clear.
2. Problems with programming
- The filters and recommended programs didn’t work properly.
- The loading speed was low.
- The site wasn’t optimized for mobile view.
What has been done
It is a part-time project which has taken 150 hours so far. What I’ve done during this time:
- Customer development (interviewing, session recordings analysis, competitor’s analysis)
- Creating a design system, prototyping and redesign.
- Finding developers to decide about the technology stack, then working with the stack including code review and testing.
- Finding SEO specialist/analyst and setting analytics with him.
- Effective and cheap visualization of complex pricing for B2B.
- Creating a landing page.
- Writing and editing the texts in the project.
- Designing adaptive and mobile versions.
Working process
Time & material instead of fixed price
I estimated the tasks and agreed upon them with the CEO. If a task seemed to take too much time I suggested a simpler and cheaper solution or explained why that task should be done in this way and why it was essential for the whole project. Then in the next column, I put real hours spent on the task, sometimes it might take a bit more time, sometimes less. It gave us flexibility and accelerated performance.
We also agreed to add an additional 15% for discussions and calls.
During the whole year, we didn’t have any arguments, misunderstandings or doubts of honesty because of an extremely clear process.
Finding developers
Our service was made on CMS platform for internet marketers. There were two main options:
- Stay on the same CMS CS Card and modify the engine.
- Choose the newer and more modern CMS Laravel and do everything from the beginning.
I sent the design to different web studios for estimation and also to some developers for independent advice. I got feedback and we chose the first option.
How I looked for the right developers:
- I checked their real products.
- I tried to find real reviews but I failed.
- We started to work with a small task to see the developers’ performance, deadlines, discussion process.
We finally found a good team but before that we had failed twice.
Challenge #1: dealing with freelance developers
The first team (developer + project manager) didn’t finish the task within the paid hours. Then they started to cheat using quite a common trick, telling us about imaginary limitations to reduce the number of tasks. They also missed all deadlines. After an independent code review, I convinced the CEO to stop working with them as soon as possible so as to save time and money.
Challenge #2: dealing with the top Upwork developer
I hoped that a new developer from Upwork with huge experience and excellent reviews could get better results. And it looked better in the beginning. But when he started to miss the deadlines we communicated with him tactfully, explaining why that was important for us, he seemingly agreed and then missed the deadlines again. His performance was extremely low and we stopped working with him after the first sprint.
Challenge #3: a sharp fall in conversion
After one of the releases, the conversion sharply decreased. The CEO wanted to try and change the color of buttons but I didn’t find it a good strategy because the probability to find the right color for even a slight increase in conversion was low, and we would waste time instead of finding the real reason. We analyzed session recordings and found the bug: it was the order form which didn’t work properly in several web browsers.
Set the ball rolling: finding good developers and sending them the design tasks
Finally, we managed to find a good team of developers. Actually, they were those who could explain their decisions more clearly and suggest different variants. I started to prepare the design for them. At first I used Sketch+Zeplin, but then moved all projects to Figma and it became more convenient to use and less time taking, e.g., when creating the components and color & text styles. To keep versioning I created new pages for each new sprint within the same project. As for the tasks, the bug list and specifications, we used Notion.
Text writing and editing
I tried to make the text shorter and clearer. I and the CEO also analyzed the top questions from the users and wrote answers to them in FAQ and tips.
Analytics and SEO
Someone was needed to do optimization in the project and I was recommended a good specialist. He agreed to help us and together we adjusted the goals in analytics and then started to work on SEO, which helped me to learn a lot of new things about it.
Not doing design for the sake of design
Although I am a designer, I prefer not to use visual design when a simpler and cheaper solution might effectively address the issue.
The company needed a way to communicate with B2B.
I suggested 2 options:
- Creating an additional user flow for them with a list of programs and a contact form. It was a long and expensive task.
- Creating a Google Spreadsheet with the options calculator and sending it to B2B.
We chose the latter and the users were able to calculate the tariff by themselves, without our spending hours on programming and designing web calculator. Customers liked our spreadsheet a lot although it looked complicated at first sight.
If you want to upload your own picture to Google Spreadsheet, I don’t recommend using Google Drive or adding them directly because on some browsers the picture doesn’t appear at all. I uploaded pictures to the social network and added a link to the cell. That worked properly.
Design process
Customer development
The good thing was that the company communicated with customers quite often. But all of their comments were from the users who had already bought the program — classical Survivorship bias.
Session recordings helped us to reduce misunderstanding about the users who decided not to buy.
Challenge #4: typical icons or modern gifs?
Most of the visitors were between 20 and 30 y.o. so I had the idea to use gifs from GIPHY instead of standard icons: it was easier than creating illustrations and looked more emotional and involving for the project audience. But there was one thing missing in web analytics: information about real customers. When I asked the CEO about them I was surprised to hear their age was 40+, over 10 years older than most of the visitors. The visitors preferred to use the service just out of curiosity, they still didn’t have chronical diseases and no real need to buy.
Customers and visitors might be from completely different segments of audience. It depends on the type of project.
We tested our two versions and realized that the users neither understood nor trusted the gifs on the landing page. We then opted for standard pictograms.
Challenge #5: «We had 500 clinics, 150 insurance programs, 3–5 sets of options…»
The service included 500 clinics, 150 insurance programs with 3–5 sets of options in each program from different insurance companies. One of the most complicated tasks was to unify the design for all of them and make them clear and convenient to use. My previous experience in redesigning Fabuza helped me a lot. I also found inspiration in modern huge online markets.
I created the first redesign. CMS platform had a lot of limitations, and it was quite expensive to design Custom Modules in CMS. Because of that, I tried to improve the current version and change it as little as possible. As a result, I missed several opportunities to improve UX. I’d used «thinking out of the box» technique before but I didn’t notice it in that case. I realized later that I ought to have started from designing a concept the way it should look best without limitations and then introduce them. Then I began to use this technique in my work process and I achieved better results.
When working with a very limited system it might be better to start from redesigning the initial concept and then change it according to limitations. If you don’t do that, you are likely to waste time on turning over various options until you finally decide to rise to a higher level and to change the concept.
Creating the landing page
Making it adaptive and mobile in 1 day
Programs were extremely heavy in terms of both UX and text. To fix that I added middle states make the pages light according to iOS guidelines and reduced the size yet keeping it legible.
It took 1 working day to adapt the screens from the web version to the mobile flow.
Current results
What we have released
- New landing page.
- New insurance program page with tips, FAQ and our service advantages, because many customers used to get directly to insurance program pages and didn’t know anything about the service.
What we haven’t released yet
- Mobile and adaptive version.
- List of programs with a new filter.
- Interactive calculator showing the advantages of using medical insurance in comparison with personal visiting.
- Real cases of usage in reviews. Although the health theme is too personal, there are some tricks of how to make it that more like success stories of healthcare and money saving.
- Loading speed optimization.
- SEO optimization.
My personal and professional improvements
- Managing skills and proactive behavior.
- Negotiating with developers.
- Responsibility for the whole product, not only for my part.
- Creating a design system.
- Creating UX of big informal pages in one style.
- Improving visual skills.
P.S. If you have any questions regarding the story, feel free to contact me via Facebook or Linkedin.
This story is published in The Startup, Medium’s largest entrepreneurship publication followed by +420,678 people.
Subscribe to receive our top stories here.
