I’ve been constantly hearing that I could be more analytical throughout the last five years of my job. Although I’ve always been confident handling and manipulating data, I believe it’s been challenging to translate that data into useful ideas. Next week, I’m switching careers to become a product manager, so I wanted to get some insights from you guys on how to become more proficient in this field.
Any suggestions are welcome?
The key, in my opinion, is to use statistics to support your conclusions.
We’re taking this action because it will lengthen usage time by Y%.
Our user discovery revealed that 60% of our consumers value B more than A, although we now spend 80% of our engineering time on A. As a result, we will increase our investment in B.
I must admit that there are situations when data is not required to get conclusions that are blatantly clear, but providing evidence increases its credibility significantly. When you provide numbers, people take you more seriously.
Data manipulation is one thing, but the first question you should ask after collecting data or insight is, “So What?”
What does it matter if “users clicked this button 5000 times”? The surface-level findings you’re undoubtedly generating need to be probed deeper.
Could you clarify what you mean by “more analytical”?
Both logical analysis and “data driven” analytics are required at certain points in time. Online, there is a ton of information about the former but less about the latter. The latter is when you use factors like recognized customer behavior patterns, organizational goals, and simply the way that some things work to rationalize your reasoning.
The next time you write a proposal, look it over and make sure everything makes sense. There are no established facts; everything could be false. Start by identifying the most dubious claims and providing evidence for them. Then, add your facts to your plan and revise it in light of what you discovered.
Spending all of your time seeking to disprove yourself is a sign that you have mastered the art of work not done. There are occasions when you’ll realize you made a mistake and shouldn’t complete the task. Sometimes you’ll be able to articulate how the work will effect your North Star Metric and zero down on which aspects of it will matter the most because you’ll have already shown to yourself that it will.
I would say data driven analytics. I’m also looking for book recommendations if you have any!
People telling you to be more analytical without providing examples smacks of poor management, so I would specifically inquire what they mean.
Since “be more analytical” typically sounds like political BS without a deeper interaction, you should immediately obtain something actionable.
What it really comes down to, in my opinion, is being overly skeptical. I’ve consistently been told that I have excellent analytical skills, despite the fact that sometimes I feel like I could do better with data. But I’ve always had a really inquisitive and skeptic attitude. Just keep asking “why” until the response is supported by a number. To a certain extent, every firm makes choices that call for a certain level of intuition and art, and it’s through these decisions that I receive my constructive criticism.
Welcome back @RobMartin, It’s been a long time, you’ve been around. Hope all’s well with you.
Coming to the original post, here’s my two cents:
Becoming more analytical as a product manager involves developing a mindset and skill set that allows you to derive actionable insights from data. Here are some tips to help you upskill in this area:
Remember that becoming more analytical is a continuous learning process. Embrace a growth mindset, be persistent, and practice regularly. Over time, you will become more adept at deriving actionable insights from data, which will greatly benefit your career as a product manager.
Treat your upcoming project proposal like a college research paper, if just as an exercise. Links to inquiries from whichever business intelligence platform your organization employs should be used to support all of your claims. Reference your sources.
That should force you to become more proficient with the toolchain used by your organization and give you the reflexes to constantly be looking for data.
It’s not usually the end of the world if you can’t refer to particular proof to support a claim, but you should be aware of it. Understand and be clear about the risks you’re taking (or asking your audience to accept).
I find that after doing that a few times, people just assume that if I’m saying something, it must be credibly anchored in data. Of course, that’s overkill in many contexts, and few people will actually click through to check your research.
Bonus points: Since leaders are the ones who will most likely examine your citations, you have the opportunity to win their favor on a budget.
@KaneMorgan, This. All of my suggestions follow the style of a high school or college debate: a recommendation, a justification for why the current situation is flawed, my solution, and its benefits.
And disadvantages I hope. No proposal of any consequence comes without trade-offs being made.
The traditional method of risk analysis. We might make 10% more money, but we’ll annoy 4% of the current user base. That seems like a reasonable compromise, but perhaps those repeat customers who are fewer are your most devoted ones. Is it worth alienating a devoted, lower-paying, but longer-lasting userbase in order to make quick money from one-time consumers who are unlikely to come back?
(There is, of course, no right answer; the one that applies to your organization, product, market, business’s financial state, and its strategic goals is the only one that matters.)
I have been receiving the same feedback as well. Learned more SQL, but it didn’t change this feedback
Learning more SQL isn’t making you more analytical - it just means you’re better at SQL.
Being more analytical means you’re able to do better analysis. You need to generate more in-depth analysis and less surface level takes. Think harder and deeper about what the data is telling you.
Can you provide a convincing instance of the difference between a surface-level and more in-depth analysis of a case?
I believed I had dug deeply into some of the data and research, but my prior employer was pushing for fresh concepts, fixes, and upgrades on a weekly basis, and we eventually reached a brick wall. The crew just couldn’t take a break because the A/B tests would only produce increases of 0.1% at most.
So, I work with a chatbot. Surface level would be “the customer portal generates 200 chats a month about customer card”. It seemed like something easily automated to order new cards as customers are authenticated and I was going to spec up a solution for automated ordering of new cards.
However, when looking at two things: 1-the amount of chats that then go through to a customer service agent was about 30 a month 2-the amount of chats themselves that automation would have helped was 2 a month
The reasons the customers needed to go through are because they need the help of a human. It’s about assessing the next level below what the data told me. Most of the 200 chats are self-contained and the small amount that go through are not worth building a solution for.
How can you dive deeper into your data? It depends on your data! It was not an onerous amounts of chats to sift through a few months worth when constraining to customer cards that I needed to grab a sample to come to those deeper level conclusions, but samples can be good when there are thousands of chats. I am very much looking forward to security okaying GPT to do that though for me though 
How does that sound and what is your product? We might be able to brainstorm ways to go to the next level!