A marketplace search flow experiment in a startup that was trying to prove himself in the Brazilian language education market.
Company: querobolsa.com.br/idiomas
Time: January to October 2019
Designer: Doug Lazarini
A marketplace search flow experiment in a startup that was trying to prove himself in the Brazilian language education market.
Company: querobolsa.com.br/idiomas
Time: January to October 2019
Designer: Doug Lazarini
My role, in this case, was interaction design. How to improve the user experience in a marketplace that offers language courses. It was a few months since I started to working at the team with design and marketing initiatives like SEO and Growth Hacking. This was my first UX experiment opportunity at the Quero team.
My role as a
UX Designer at Quero Languages
My role, in this case, was interaction design. How to improve the user experience in a marketplace that offers language courses. It was a few months since I started to working at the team with design and marketing initiatives like SEO and Growth Hacking. This was my first UX experiment opportunity at the Quero team.
Quero Education is the company responsible for some of the largest educational marketplaces in Brazil, offering discounts on higher education, technical courses, language courses, and even elementary school.
I has hired to work in a business vertical of new products, after their louched the first new product was Quero Languages marketplace, the objective was reproduce the success that the company had in higher education. The problem we face was the differences at the market, like courses peculiarities and characteristics that were not considered in the initial phase. Due to these problems, users did not find the courses they wanted and the self-service sales rate was very low, most sales needed call assistance for help to buy and they were already overwhelmed.
Context and scenario
Quero Education is the company responsible for some of the largest educational marketplaces in Brazil, offering discounts on higher education, technical courses, language courses, and even elementary school.
I has hired to work in a business vertical of new products, after their louched the first new product was Quero Languages marketplace, the objective was reproduce the success that the company had in higher education. The problem we face was the differences at the market, like courses peculiarities and characteristics that were not considered in the initial phase. Due to these problems, users did not find the courses they wanted and the self-service sales rate was very low, most sales needed call assistance for help to buy and they were already overwhelmed.
My first action was to understand the higher education user, and to my surprise, we only had one persona and one documented journey, the company had no experience in qualitative research. We were trying to sell language courses to users looking for higher courses on the Quero Bolsa marketplace and we had no data on who our audience really was. I looked for more information with desk researches to understand the main reasons and profiles that lead users to acquire language courses.
2. Analyzing user metrics
With the Google Analytics data, I could understand who was looking for our services. The platform stored demographic data, we filtered who was attracted organically (without ads) to our marketplace and this helped us to have some ideas that would be a little clearer ahead.
3. Phone interview with Business Partners
One of the ways to bring a little bit more information about our users was by talking with language school owners about customers’ key needs, latent needs, and how they react to the scarcity of discount coupons.
4. Guerrilla and Shadowing
One way I found to understand users was by participating in experimental classes when it was the first class was even better, as students shared the reasons for doing English. Some responses were: “Because my mother forced me!”, “I want to immigrate” and “To improve my professional qualification!”.
5. Using Hotjar to map the user flow
We used Hotjar to record and observe several people trying to find a language course and discovered the same usability problems in the search flow. We name it “Offer’s Wall” because we displayed so many similar offers that the user did not know what the difference was between them and continue to roll for two or three pages until contacting the chat or calling for help.
User understanding
1. Desk Research
My first action was to understand the higher education user, and to my surprise, we only had one persona and one documented journey, the company had no experience in qualitative research. We were trying to sell language courses to users looking for higher courses on the Quero Bolsa marketplace and we had no data on who our audience really was. I looked for more information with desk researches to understand the main reasons and profiles that lead users to acquire language courses.
2. Analyzing user metrics
With the Google Analytics data, I could understand who was looking for our services. The platform stored demographic data, we filtered who was attracted organically (without ads) to our marketplace and this helped us to have some ideas that would be a little clearer ahead.
3. Phone interview with Business Partners
One of the ways to bring a little bit more information about our users was by talking with language school owners about customers’ key needs, latent needs, and how they react to the scarcity of discount coupons.
4. Guerrilla and Shadowing
One way I found to understand users was by participating in experimental classes when it was the first class was even better, as students shared the reasons for doing English. Some responses were: “Because my mother forced me!”, “I want to immigrate” and “To improve my professional qualification!”.
5. Using Hotjar to map the user flow
We used Hotjar to record and observe several people trying to find a language course and discovered the same usability problems in the search flow. We name it “Offer’s Wall” because we displayed so many similar offers that the user did not know what the difference was between them and continue to roll for two or three pages until contacting the chat or calling for help.
Solutions
After analyzing many users who tried to use the search flow and failed, we identified problematic flow patterns. What we identified was that schools with classroom courses (offline) had a very large number of course variations and each of those was a different offer (SKU). We would have to deal with those variations so we supose that the volume of courses on the search page will decrease and make it easier to find the course they want.
Solution proposals
To show the characteristics and differences on the product page itself, the initial idea was to create generic offers in the survey and allow users to select the characteristics on the offer details page. An example of this practice is the clothing e-commerces, where the user needs to select colors and sizes. In our case, the user would have to select age, level, and time. This solution was discarded mainly due to technical difficulties, it involved complexities that we could not meet at that time.
After analyzing all the constrains, we presented a proposal for something that directly involved the listing of offers, a way to group dozens of offers into a single card. We find something similar in Booking.com, other teams was testing something similar, and we also want to test that. The offers was grouped by schools, classes ages, levels and classes time.
Solutions
After analyzing many users who tried to use the search flow and failed, we identified problematic flow patterns. What we identified was that schools with classroom courses (offline) had a very large number of course variations and each of those was a different offer (SKU). We would have to deal with those variations so we supose that the volume of courses on the search page will decrease and make it easier to find the course they want.
Solution proposals
To show the characteristics and differences on the product page itself, the initial idea was to create generic offers in the survey and allow users to select the characteristics on the offer details page. An example of this practice is the clothing e-commerces, where the user needs to select colors and sizes. In our case, the user would have to select age, level, and time. This solution was discarded mainly due to technical difficulties, it involved complexities that we could not meet at that time.
After analyzing all the constrains, we presented a proposal for something that directly involved the listing of offers, a way to group dozens of offers into a single card. We find something similar in Booking.com, other teams was testing something similar, and we also want to test that. The offers was grouped by schools, classes ages, levels and classes time.
Design solution
After validate all commercial and technical constrains, we started to develop the solution, which was a user flow, interactive prototype, and usability test.
I started mapping the user flow, so would be possible to present the idea to the team and stakeholders.
The prototype was developed in Sketch and Abstract using the company’s Design System.
It was not part of the company’s culture to do qualitative research, they preferred to test with metrics of use in production, once again we had to improvise. When the solution entered the approval phase, we carried out usability tests with employees of the company, that we identified some problems of taxonomy and usability. We didn’t do the usability test and taking the next stage of the feature, but we was going to follow with metrics.
Design solution
After validate all commercial and technical constrains, we started to develop the solution, which was a user flow, interactive prototype, and usability test.
I started mapping the user flow, so it would be possible to present to stakeholders what the idea was, to do the first validation.
The prototype was developed in Sketch and Abstract using the company’s Design System.
It was not part of the company’s culture to do qualitative research, they preferred to test with metrics of use in production, once again we had to improvise. When the solution entered the approval phase, we carried out usability tests with employees of the company, that we identified some problems of taxonomy and usability. We didn’t do the usability test and taking the next stage of the feature, but we was going to follow with metrics.
Results
Achieved objectives
Different objectives were determined and changed during the semester, and we had a little difficult to understand what would really bring value to the company. In the end, we were with just two main objectives, one for the product and another about the business vertical.
Increase of 30% in self-service conversions and consequently decrease in the volume of phone and chat contacts;
This experiment, as well as actions in other sectors such as marketing and sales, contributed to the company reaching the billing goal, and the business vertical break-even.
Results
Achieved objectives
Different objectives were determined and changed during the semester, and we had a little difficult to understand what would really bring value to the company. In the end, we were with just two main objectives, one for the product and another about the business vertical.
Increase of 30% in self-service conversions and consequently decrease in the volume of phone and chat contacts;
This experiment, as well as actions in other sectors such as marketing and sales, contributed to the company reaching the billing goal, and the business vertical break-even.
Learnings
At the end of semester the company aquired another company specialized on basic education, the entire task force of the business vertical going to focus on that new operation, so that the language marketplace switch to passive mode.
That was sad but we were able to learn good lessons from all that. That’s why we do experiments, not only to validate or invalidate hypotheses but mainly to learn from them:
1. “We didn’t have we learned to replicate the success we had in the higher education marketplace”. Working with a digital product is extremely complex, if we don’t align values and expectations frequently, we lose control of what is really important for the company and our users.
2. After this case, I left the company to face new challenges. If I could choose a single lesson to take with me it would be “It is not enough to be a good designer, to understand interaction, visual design, design system, and UX research, it is necessary to understand the product, business, and especially people.” The designer is an extension of the product, he needs to work in a collaborative way, generate inputs, align expectations, and show the best way forward, that was the best lesson I could learn from Quero Education.
Thank you to get this far. It looks simple in the portfolio, but it was a difficult and incredible professional challenge.
If you want to talk about that or about other things, please contact me.
Contacts
+55 (12) 99181-3891
linkedin.com/in/lazarini/
Learnings
At the end of semester the company aquired another company specialized on basic education, the entire task force of the business vertical going to focus on that new operation, so that the language marketplace switch to passive mode.
That was sad but we were able to learn good lessons from all that. That’s why we do experiments, not only to validate or invalidate hypotheses but mainly to learn from them:
1. “We didn’t have we learned to replicate the success we had in the higher education marketplace”. Working with a digital product is extremely complex, if we don’t align values and expectations frequently, we lose control of what is really important for the company and our users.
2. After this case, I left the company to face new challenges. If I could choose a single lesson to take with me it would be “It is not enough to be a good designer, to understand interaction, visual design, design system, and UX research, it is necessary to understand the product, business, and especially people.” The designer is an extension of the product, he needs to work in a collaborative way, generate inputs, align expectations, and show the best way forward, that was the best lesson I could learn from Quero Education.
Thank you to get this far. It looks simple in the portfolio, but it was a difficult and incredible professional challenge.
If you want to talk about that or about other things, please contact me.
Contacts
+55 (12) 99181-3891
linkedin.com/in/lazarini/
