Diego Perdana – Product Manager Portofolio

This post is my reflection for my previous job as Product Manager in Discovery and specialize on the Search engine product. Here’s what I got and notes for me

Growth the Product and Zero to One Product

Join in the new team of Search, I already manage the existing product from Global Search bar until the Search experiences. The Search product it self is one of the core features on marketplace like Tokopedia with a really huge traffic and facing buyer journey to find product easier and faster. Also, we have a big OKR for this search to maintain the performance including CVR and CTR. From this team i learn to become a PM that focus to growing the existing product to meet the OKR target each month by improve the current features with small or big improvements. I learn a lot is about the funneling method to breakdown the current metrics and improve at the broken or unoptimized metrics with small improvement like in ranking, logic or event small UI.

Recently I move to the new team that focus on building a new Product. The product is creating Search as a service platform like Algolia.com to helping tribes Tokopedia building the search feature. When handling this product, its about 0 to 1 product, I learn about how we building the real MVP of 1 product and new business. I need to drive the vision, thinking about the business, roadmap for 2 years and also approaching the clients. But I think I also made a mistake when creating this. The mistake is we focus a lot on the clients and not prioritize the platform product itself. We cover up a lot clients need by approaching one by one and do the development to prove our concept. Meanwhile there are another approach when we figure out all the clients needs first, put into one bucket and create the product based on majority client problem. But again its depend on the situation and manpower, because when building this we lack of expert and people who cover up this. That’s oke i learn a lot from that.

A/B Testing experiment is required

In Discovery platform especially on Search and Recommendation who have a big traffic and lots of users personas we should know when and how to release new product or improvement. By then the usual things that we should use is A/B testing experiment. You can search what is A/B testing experiment on Google to find the meaningful explanation.

In Discovery it self we do a lots of A/B testing experiment or gradual roll out to checking the feature. The goal for this A/B testing experiment is to check our solution is impactful enough to our users based on metrics that we want to increase (most likely is north star metrics) without release it to all users. So we now that our solution is fits or not with users.

Whitelisted VS Blacklisted approach

Discovery tribe is about handling a many use cases from other business because we are a platform. There are also so many approach especially as a platform we need to maintain the relevancy. Two approach if we are want to add the feature on production or even though A/B testing we need to make sure we know how to use whitelisted or blacklisted.

Whitelisted means you allow some rules to be open first one by one, like the feature only whitelist to certain users or open for certain case. This approach is really good for maintaining the current metrics when released the new or modify the features. For example in Relevancy side on search, we can release one widget in one group of keywords

Blacklisted means we are force to not showing this to certain criteria, like the feature is cannot be show in environment apps iOS or certain keywords need to be blacklisted so users cannot see it.

Unification create scaling feature

As in my Search team I able to learn about the unification project to reducing complexity but scale a lots. This project is really importance for product that have a lots of component to testing or added. For example:

Unification of tracking to create a simple tracking to be maintain but scalable if there are a lots of component to test. We just need to maintain only 5 tracking pattern but because the value or configurable things coming from BE so we can control it without having a FE released.

Unification in design component to create a unify component that can be used as Global component to reduce the development effort from FE and already implemented the existing field tracking. For example the Search Result Page layout in Global Search is same in other page because we use the global component of the layout.

As a Platform we should handle the specific problem into scaling solution

Search is a platform product where all the business and problem coming from outside not only inside (what i mean is PM and team inside Search). By then we should gather the problem from one users and talk to another one if they find the same problem. For example, if one category report that they need one feature on Search, we should check another category team if they have same problem. By then we should create the solution that can be work both of category and can be used for another use case.

Writing to define the customer problem

Recently i’m just realize that this is the most important things that we should do and deep dive to the real customers if they have a problem or not. Also we need to articulate what the problems is into writing document so we know that if its the problem exist or not. The writing document i got is based on Amazon expert when they want building one product.

Even though we know the customer problem based on user experience or NPS in our mind, when we write it on document somehow is difficult but it’s good exercise. Because you need to think when your writing we need to make sure our document is understand well to our stakeholders so they know the real problem and how big it is.

3 Value: Focus on Consumer, Make it Happen and Make it Better, and Growth Mindset

Those 3 values on the section title is Tokopedia DNA and i really want to keep it when i create the products. Those values are applicable when i building product, for example, when i building the cross selling widget on search, we focus on the customer problem by finding evidence though data and talking with our customers (Nakama and Buyer). After that, we together create the MVP of product which is to make it happen first. After the product release we got a lots of feedback about the performance and system itself that make us need to make it better the product. We taken the feedback even though is negative from customers and data to find it the best improvement and that’s the part of growth mindset