The Journey of Setting Up Continuous Discovery Process at GOGOX

Lili Chan
GOGOX Technology
Published in
8 min readOct 21, 2020

Ask any good product teams and they’ll say that product discovery is critical in their product decision making process. However, in reality, many product teams are so busy with their delivery tasks and are not able to spend enough time to discover the right opportunities and solutions to work on.

At GOGOX, we had similar problems in the past as we grew from a local van hailing app in Hong Kong to one of the top logistics platforms in APAC. It became a challenge for the product team at our headquarter to understand our users across countries on a continuous basis. At first we relied on occasional product research trips, data analytics and user feedback from different channels to solve this problem. But then we found that the findings often became outdated rather quickly. We hoped to find a more systematic approach that would help us move fast and be more customer-centered — that’s when Continuous Discovery came into the picture.

Product experts like Marty Cagan and Teresa Torres have long been advocates of Continuous Discovery. It is a process of dual-track development helping the team strike a balance between their short term and long term goals. Product teams are able to evolve and refine their ideas based on customers’ needs by conducting small research activities (e.g. interviewing, prototyping, experimenting) with customers on a frequent basis.

The process may sound trivial but it’s not without its complications, especially when our goal is to help teams gain valuable insights and integrate them into future product iterations.

In this article, we want to share some of what we’ve learned from helping GOGOX gradually adopt the Continuous Discovery process.

Research Challenges in a Global 2-sided Marketplace

As a result of rapidly expanding our business in the past 7 years, our markets are at different stages of maturity. This means we have to tweak our research topics depending on the market situation. In the past, the product team and local operations team conducted research separately which resulted in overlapped findings. Apart from that, findings were not proactively shared to a wider audience so there was no single source of truth — it would be buried deeply in email threads, get scattered somewhere on Google Drive, or become lost in Notion pages.

In Asia, the logistics industry is more traditional and scattered; clients typically rely on calling service hotlines to find a driver. In a 2-sided marketplace like GOGOX, while the process can be done easily through an app, clients and drivers do have different expectations. For example:

  • Clients will write order details in the remarks expecting the drivers will read them
  • Drivers seldom read the order remarks because they think they could understand the request through a phone call

Additionally, since the age range of our users is broad, the older groups usually don’t know about our app features and stick to minimal functionality. For younger age groups, they are more tech savvy to things that can provide them extra convenience. It is challenging for us to strike a balance between the demographic extremes.

Using Continuous Discovery to Solve Our Problems

In order to solve the above challenges, we’ve set up a CD (Continuous Discovery) team consisting of our product design lead Gemmy Wong and one product manager. The CD team works with product designers to create and maintain a research playbook that consists of research method tutorials along with a generic set of questions related to our user journey.

After identifying top priority problems that PMs and designers would like to solve in the upcoming quarter, research topics are incorporated into the interview question list. We ask the country teams to help translate the questions and materials into local languages. At the same time, we work with the data team to obtain user lists that suit our research. The CD team is responsible for user recruitment with the help of local teams. Once the interviewee list is confirmed, the local teams will facilitate the interviews and document the findings. Then they will be passed to designers to consolidate the results, and update the user journey map. At the end, a research insights report/sharing is prepared by the CD team and shared with stakeholders.

At the moment, we are using Notion as the single source of truth. Question sets, personas, research topics, schedules etc. are all found there. So different teams can easily find the research materials and results when needed.

Lessons Learned So Far

One thing that caught us off guard was the time & effort required for participant recruitment. As we have a smaller user base in newer markets, we would often have to rely on calling users individually to secure participants if our message engagement rate was low. We sometimes overestimated the client’s willingness to share their feedback with us. And since GOGOX is not an everyday app like Uber or Lyft, users may not be familiar with our app enough to give their thoughts on it.

The current COVID situation has also changed the way we conducted interviews since local team members couldn’t meet our users in person — this introduced remote usability testing as a new challenge for us to tackle. We understood that not all users were tech savvy enough to use video calling apps for screen sharing, and we typically conduct our interviews during working hours, which might not have always made it convenient for clients to have a long call with use over the phone or share screen in their workspace.

We also faced issues surrounding usability testing. Usability testing can sometimes involve a prototype — which would require a clear understanding of the flow. Typically, we had designers conduct usability tests with business team members as the facilitators. After our first round of usability testing was completed, we found out that some of our business team members were not too familiar with our product UX. This made it a bit tricky for them to carry out usability tests themselves. Quick tip: Local team members can also be good interviewees for your usability testing!

After doing this for a few quarters, we have realized it’s important to set the right expectations with the business team members on interview content quality — like when to ask follow up questions instead of simply asking what’s on the question list. It actually takes time for team members to learn how to conduct research, so be sure to show empathy to local team members who help you with executing the task because they could be a first-timer.

Here are a few tips that may help if your team is considering adopting Continuous Discovery:

For a CD team…

  • Write a clear user manual. Assume this is a suitcase that can be passed to anyone — use simple language.
  • Walkthrough the process with interview facilitators. Consider this is a trial run for the interviews, and you will often find room for improvement.
  • Involve business team members at an early stage when creating a localized version of interview questions. Gather feedback on potential blind spots since they are the one who know best about their respective markets.
  • Make good use of data as a reference, or use it as a starting point to gather question ideas.
    ・E.g. We found the ratio to rate an order is low and we wanted to focus on service quality in the upcoming quarter — questions regarding intention to rate an order would be insightful.
  • Stating the details of your research in the invitation message would help to increase the conversion rate
    ・What is the interview format? How long is the duration? Any incentives?
    ・ Personalize the invitation — Assuming the CD team doesn’t involve a marketer in this process, personalization is your key to drive engagement! Incentives can be very eye catching!
  • Optimise the length of the interview and the amount of incentive. Within 1 hour would be ideal — subject to COVID outbreak remote interviews should be shorter than that.

For interviewers…

  • Be curious about the user’s response. Even if it’s just a single yes/no question, you can ask them why.
  • Notice the non-verbal gestures of interviewees. Verbal expression can be implicit but you can observe their gestures when interacting with your product.
  • Translate, but don’t interpret. We all perceive things differently, which could affect our findings if we add in subjective thoughts there.
    “It’s quite easy to use” VS interpreting the user had no issue with using the app
  • Don’t ask leading questions
    “Do you think this button is obvious enough?”

How Our Continuous Discovery Process Will Evolve

As a global logistics platform, we may not be able to talk with our users in person as frequently as we want. To make the continuous discovery process sustainable in GOGOX, we will be exploring cost efficient ways to gather user signals. This includes making recurring interviews smaller in scale — we will work with the engineering team to build in-app surveys that we can set up for newly iterated and launched features. Overall resource cost will be reduced, and we will be able to gather more quantitative data.

We will also be gradually scaling up our efforts to include signals from different channels. Users now reach out to us in different ways and we haven’t been able to address them all at once; with the help from cross functional team members we hope to be more “customer first” than ever before. A few things that we need to do are to automate the signal collection process, and involve the data team in centralising the feedback analysis — thus unifying the signal categories across cross functional teams.

With all the signals categorized and stacked, we will then be able to backtrack and analyze the qualitative data against quantitative metrics. For example, verifying user pain points with our data to validate whether an issue is critically affecting the conversion funnel. We are still in in the midst of shaping the continuous discovery process, but it is already helping us in product enhancement:

We found out that users didn’t expect that tapping the “back” button on the vehicle screen will discard the order. We’ve added a dialog to warn users about that.
Users are not aware of the white add waypoint button. We make it more prominent by changing it to our primary blue colour.

Enabling continuous discovery is a challenging process for product teams, but the effort spent is definitely worthwhile to help your company become more customer oriented. Feel free to reach us at research@gogox.com if you want to talk more about continuous discovery process!

Thanks to my manager Vincent Chan and design lead Gemmy Wong for making this happen! Also Leslie, Waikit, Crystal, Tom Lui, Phoenix, Rob K, Tom Lau, Matilda, Cupi, Franco, Jai, Alvin, Fisher, Alvan, Hao Feng & Emme🙏🏻

Last but not least — we are hiring new talent to join our teams! Check out our careers page for more details.

--

--