Top product management reads in October

4 min read
Share on

From surveys on product tools to burnout and building AI products that aren’t terrible, here’s a roundup of the most-read posts on Mind the Product this month.

Our most widely read article this month laid out the findings from some polls we conducted to find out what tools the product community have found most valuable to use over the past 12 months. Pendo, Jira Product Discovery and Perplexity proved to be the most popular.

Poll results: What tools are product managers using the most?

This post from Sapnil Bhatnagar looks at how resistance from stakeholders can often become a significant roadblock to delivery, innovation and documenting progress. 

He runs through some key themes of stakeholder resistance, namely: vision misalignment; risk aversion; communication barriers; legacy system dependence; budget constraints; lack of trust. He also looks at what in his experience has worked to overcome each type of resistance. For example he used OKRs to align the product vision and achieved a 40% reduction in feature requests. He comments: “With the right strategies and mindset, it's entirely possible to turn sceptics into advocates. The first step is to embrace resistance as an opportunity to improve, and iteratively optimise one step at a time.”

Leading B2B product management in high-resistance environments

This post summarises some of the best advice on avoiding and dealing with burnout. In short you should try to: set clear boundaries; prioritise and delegate; communicate openly; practise self care; set realistic goals; enlist peer support; build a resilient mindset; be aware of and assess your workplace culture; make time for personal creativity; don’t be shy of seeking professional help. It’s important to remember that taking care of your mental health not only benefits you as an individual but also leads to more effective decision-making, creativity, and long-term success in the role.

Burnout in product managers: A World Mental Health Day report

Another survey post this month set out to evaluate how features are received by users. It found that the average feature adoption rate for products is 6.4% and that manufacturing and consumer goods products have the highest feature adoption rate. And companies with fewer than 200 employees have the highest feature adoption.

The post also looks at ways to improve feature adoption - including simplifying with walk-throughs, targeted communication and announcing new features in-app.

Users engage with only 6% of product features: Product benchmark findings

In this post, Archana Kumari starts with the premise that most AI products fail to align with user expectations, lack real-world adaptability, or are a poor use case for AI and then looks at how to build a successful AI product. She uses spam detection as an example of how to use AI to build successful products

Without clear problem definition even the most sophisticated AI models won’t deliver value because they won’t align with users' needs. Does spam mean any unsolicited message, for example? She looks at how the problem of spam detection has been solved traditionally through rules and at the introduction of machine learning and adaptive filtering.

Large language models(LLMs) are a game changer, Archana says. But while they're super powerful, they can also be super unpredictable. They rely on the data they’re trained on, and if that data is biased or limited, they’ll make mistakes. They can be really hard to debug and are expensive to run. 

Archana then offers a blueprint for building a spam detection system. This means you should start simple, then scale, know when to use LLMs, use real-world data for training, iterate with user feedback, and that transparency is key. You should also continuously monitor and update, and make sure your systems are being used ethically and responsibly.

How to build an AI product that doesn’t suck