A Journey of Delivering Impacts
My PM Story at Tokopedia
Disclaimer: I am just a PM with a 3-year experience. This is just a story of me, doing the execution, and then when I view holistically, I realized that these tiny little pieces of skills are the ones that moved the ball forward. That is: being collaborative and data-driven.
I believe working at different startup stages requires a different approach to be a PM. A startup PM in a ‘problem/solution fit’ stage, might need to do more in UX Research to know their customers deeply, including their behavior, pains, and gains. A startup PM in a ‘product/market fit’ stage, might need to do more about validating their products, conducting market research, and also technical architecture decisions. Lastly, a startup PM in the ‘scaling’ stage like Tokopedia, needs to streamline processes, improve performance, reducing cost, and so forth. I will tell you my story about the latest stage, whose data is already so huge, but somehow didn’t catch to attention.
I just relocated from OVO squad, a squad to lead integration between Tokopedia and OVO, as per July 2019. Until then, I got the responsibility to lead the Payment Tools squad, which is a squad to help customer excellence. So, I need to make sure that consumers’ payment experience is seamless, easy, and secure. Information to users needs to be clear in all payment products and also when users try to create customer service tickets. Additionally, I need to create internal tools to help the escrow team and transaction team do their job easily with automation.
One thing that I like in Tokopedia is we work on the ‘ROWE’ environment, a Result Only Work Environment. That is why, we mainly focus on how to achieve our targets, but the ways to achieve that are just limitless, so we need to be creative on our proposed solutions. Just not so long ago, we shifted the cultural performance using OKRs (Objective and Key Results). So, one of my Key Results is Percentages of Customer Services Tickets contribution to payment tribe per total transactions. In Q1–Q2 though, we didn’t perform that well. Payment CS Tickets almost always became the top 3 for the most cs ticket contribution to tribes. For your info, on a daily basis, we monitor the same number for 16 tribes. Payment CS Tickets fluctuated between 0.009% the lowest to 0.017% the highest. The average was 0.012%. This was expected because payments are the key to Tokopedia’s transactions. After some discussions, we aim to reduce the Q3-Q4 target to 0.009% in December.
Here I was, new to the squad with a tough challenge ahead. Fortunately, I got some support from my boss. First, I was introduced to the problems, the playing field, and ultimately, data points that can be checked. Second, I was introduced to the operations tribe, which focuses on consumer resolutions, including chatbot and auto-replies. So, here are my 4-steps actions in my journey to reduce CS Tickets.
1.Understand The Problem and Stakeholders
Doing this is a challenge, at first, I thought this is a small-group job and the cases are not unavoidable. I started to work with colleagues. as Payment has our SRM team (Service Recovery), I created weekly brainstorm sessions with them to get insights from our consumers. I found out that payment had so many problem categories that we call it Join Categories. Here are a few of our biggest categories: ‘Instant Payment belum berhasil’ (instant payment not verified), ‘Pesanan batal dana belum kembali’ (Transaction canceled, payment not yet refunded), and ‘Penarikan dana belum masuk rekening’ (withdrawal pending).
That is progress, but, knowing that is not enough. I needed to understand the case holistically. Slowly, I began to understand that this is a task for a large team to really move the needle. Here I mention all of the roles, including their role to reduce CS Tickets.
A) Level 1 — Customer Service: the people who really get in touch with customers.
B) Level 2 — Transactions: the people who manually make decisions when automation is not working.
C) Level 3 — SRM: a team to inform features and problems from payment tribe to customer services. Also, the ones who propose articles
D) Customer Experience (user side): the team to understand user journey and problems.
E) Customer Experience (resolution side): a person who comes up with ideas to personalized auto-reply for cs tickets coming from the web.
F) Knowledge: a team which creates join category, maintain article mapping according to users’ payment status and order status.
G) Resolution: we call it Gandalf, which provides one time only web-reply automation when the user creates a CS ticket.
H) Chatbot: a personalized chat reply robot based on the decision tree model.
I) Salesforce: a squad to handle Salesforce system for Tokopedia.
J) Data Analyst: a team to crunch data, creates monitoring, and come up with insights.
K) Software Engineers: handles payment features.
As I am in charge of this key result, I need to lead weekly sessions with those guys to make sure that we ideate, create solutions, and drive impact. So since then, we organize a one hour sync up every Tuesday at 17.00. 15 minutes spent to present last week’s impact. Next, 30 minutes is allocated to mention problem statements and propose a few solutions. The last 15 minutes is to write down the task, PIC, and timeline.
Next, I need to get my hands dirty to check hundreds of complaints daily, to really understand what problems are the consumers are facing, and how our CS Tickets deflection system works. For example, after reading some complaints, I understand that there are payment verification problems. But then, when I deeply analyze the root cause, it turned out that the real problem is actually there is some miss calculation in coupon code/promotion, which leads to his invoice is not generated. So I need to make sure that agents know this case, auto-reply works well, API refund/cancel from order works well, CRON running periodically, and so on.
2.Proposed Solution
We needed to have a framework to work collaboratively, at a fast pace. One night, I desperately looking for a solution to address a few problems, namely:
- How might we* monitor our performance? (Since looking at CS Tickets number only won’t take us anywhere)
- How might we* come up with an end-to-end funnel to make sure that every deflection is monitored?
- How might we* do the project management to make sure that every initiative is recorded, as well as the impact?
- How might we* know existing features and solutions in order to propose solutions?
*I brainstormed using the HMW (How Might We) exercise in Design Sprint
I got crazy until 22.00 in the night when I finally decrypt the framework that can solve points as mentioned above. I proudly presented to the team about this framework, and they found it satisfying.
After unwinding those complexities, finally, we can prioritize the things which matter most, namely those aspects which contribute the most tickets and the lowest deflections. Subsequently, we also working on navigation problems, as customers may create CS Tickets in several ways, experience in creating CS Tickets might different one another. Ultimately, every task that we did, we can monitor its impact immediately.
3.Collaboration, Commitment, Trust
Just right after entering Q3, I have just realized one thing that is hugely essential. That is the importance of collaboration. Somehow, I recognized that few new people in Tokopedia, including myself, have this arrogant tendency when we held a meeting at first. Becoming assertive and controlling might be the first option to achieve ‘that is how we get heard’ principle and showing our capabilities. Yet, that is really not the best way to work in this fast-paced environment which needs collective intelligence to obtain improvements and progressions. Instead, I needed to be a leader that has a high level of influence just like the theory from ‘Radical Candor’ book by Kim Scott.
Radical Candor is Care Personally and Challenge Directly at the same time
Apple and Google, who have a different set of cultures, show that same principle on how their bosses behave to their subordinates. Steve Jobs, for example, one time, had an employee. His employee knew he was right, but then as Steve push his own idea forward, it made his employee became silent. After a few days, Steve found that his employee’s opinion was right, so he came to his employee, and say, “this was your idea. But you are still wrong.”
“But why?”
“Because you failed to get your point across.”
Challenge people and ready to be challenged at the same time guarantees collaboration to form.
Subsequently, after the collaboration is encouraged, everybody should put their commitment to finish the tasks assigned to them. We document it in one excel sheet so everyone can revisit it later. When the iteration of collaboration and commitment is solidly formed, everybody will begin to trust each other and thus, will push their performance forward.
4.Data
Lastly, I monitor our data daily in order to know the impact of every initiative that had been implemented.
We checked the data which aligns with our CS Tickets framework. Here are a few of the examples:
- In join category dashboard, for initiative code ‘Transaction canceled, not yet refunded’, when we improve the mapping (so people will look at the right article according to their order and payment status), we improved correct navigation from 21% to 67%.
- In the chatbot deflection dashboard, for initiative code ‘payment not verified digital’, before changes, the leaf intent created 256 tickets in the first half of the month July 2019 and only created 1 ticket on the last half of the month. This resulting in a 25.500% deflection improvements.
- In the web auto-reply (Gandalf) dashboard, for initiative code ‘Refund-order rejected’ deflection rate rose from 50% to 55%.
- Ultimately, Payment CS tickets reduced 42% to only 0.007%, exceeding our target in Q4 (0.009%), and we still have 2 months to go.
I am really proud of my team who are really creative, committed, and impact-driven. Clearly, this experience is very valuable to me. I learned that becoming collaborative and data-driven is the key to deliver a profound impact as a Product Manager. I hope, this writing and framework of thinking might add value to you as well!
