An efficient way to start product discovery 🚀
4 easy steps to get into product discovery… also if you are working in a “delivery” modus company
🔍 Product discovery is a non-negotiable task for every product team. The product manager (incl. owner.. which is the same thing) should do discovery work every week.
However, if you are a young product manager (PM) in a classic corporate startup, you may find it hard to juggle product discovery with stakeholder management, building a roadmap, jumping from meetings to meetings, and finally filling up the backlog.
In this article, there will be 4 tactics that empower a PM to start product discovery in the right way. The outcome of the article is to provide confidence in the way discovery is approached so that it will be part of any PM weekly routine.
What is Product Discovery
“Product discovery refers to the activities required to determine if and why a product should be developed. Carrying out this work makes it more likely to create a product user actually wants and needs” — Roman Pichler
This is the definition of product discovery by Roman Pichler, a leader in product development (1). It basically means, finds out what needs to be built to decrease uncertainty.
Now let’s get into it.
Step 1: What is the goal, what are you trying to impact
Work against clearly defined metrics. I recommend using the North Star Metric “framework” to set a goal. To qualify as a “North Star ⭐️,” a metric must do three things: lead to revenue, reflect customer value, and measure progress (2). When this is set, it is time to plan your product discovery. Any testing should lead to an impact on the North Star Metric.
Lead to revenue 💸: A metric that leads to revenue can also be seen as a metric that makes the business viable. An example could be a KPI that creates interests in VC’s eyes.
Lead to customer value 🙃: this is crucial for the NSM. When. continuous customer value is created, there are higher chances that the customer sticks with the product longer and automatically will increase the customer lifetime value.
By using only one metric, it is easy to compare it over time and present it to colleagues and stakeholders.
Step 2: Be hypothesis-driven and structure your testing
Test hypothesis and not features. When testing hypothesis, the desired behavior-change of a user is tested. Whereas if the focus on doing discovery is for a feature, the results will lack knowledge of why it was successful or not.
At the very end of a discovery process, is good to test a full-fledged feature. Most of the time testing is focused on problems and ways to solve them.
Use the 🧪 testing card create by Strategyzer when setting up a new experiment. This helps to transform guesses into verifiable assumptions. Despite there are only 4 questions, it is a difficult task to complete.
The testing card forces us to take our assumptions (which we treat like facts) and put them on the stand, to challenge and verify/ contradict our assumptions.
Try it now! Strategyzer Card:
I use this card on most of the experiments I perform. Follow this link for a more detailed explanation on how to use it.
The main advantages are that it forces you to structure your hypothesis into quantifiable results. Try to use this format also when you ran 1:1 interviews
Try it now pt. 2! Product Academy template:
I recently participated at a Workshop by Tanja Lau and the Product Academy (best Workshop ever — > here is the website). An advanced hypothesis testing card was presented. I haven’t used it much yet, but I like it because it adds reasoning why the experiment could be successful. It adds an intrinsic emotion to the testing we are currently running. I truly believe that if you can tap into the psychological root cause of why a hypothesis is verified or not, you are a step ahead.
Step3: Who is part of your discovery?
Don’t try to validate a hypothesis with the full user base. For example, if the testing is for a copy for first-time users, there is no need of including returning users in the stats. They need to be taken out. If cohorts or segments are representing a realistic persona, then there is a powerful audience-based tracking in the company.
However, most junior PM’s out there are not yet aware of what behaviors define what persona (this is a very hard thing to understand for any organization). The solution is to think reversed 🔄. It is much easier to filter out the behavior of users that are not part of the desired testing. These patterns are normally easier to spot and therefore eliminate from the analysis.
Another way t slice the whole users base is by using the AARRR framework. This 5-steps funnel divides the users into their maturity in their product journey. This is a great way to start thinking and eliminating users that are not relevant to the current goal.
Step4: D-A-T-A
Before starting discovery, a strong data tracking setup on the product has to be installed.
Every morning check your data — Marty Cagan.
Mr. Cagan said something like that, and I couldn’t agree more.
When data 🔢 is backing up the research, the “odds” to win an argument/ topic/ idea are much higher. But unfortunately not always is enough to convince other stakeholders.
The data at the beginning of an experiment must be also available and measured. So that, there is a baseline to say “yes this is validated” or the opposite. In an old-school corporate startup, the PM might be the only one looking into the data. Therefore is even more important that the dev team is kept up to date on what is happening in the market. They work every day to solve other people’s problems. When the hypothesis is set up, make sure that the metrics to be measured are leading indicators of the success/failure of the test.
Another piece of advice I could personally give you is not to use Adobe Analytics. There are many more analytics products like Google Analytics, Amplitude, Segment, etc. which are much more intuitive and empower the PM to do his/her job better.
In conclusion, please remember the following:
With a clear goal, is easier to build the right hypothesis. Select the users who will be tested and use data to determine the success/ failure of your activity.
These steps are going to help you structure your approach into product discovery. I am excited for you to do it and remember that also a failed experiment should be celebrated.
Life is too short to build products no one wants — Ash Maruya
Thank you. 🙏
Tiziano Nessi
