Ben Thompson (via his Stratechery brand) is unquestionably the most influential commentator on Product Strategy right now and Aggregation Theory is his Relativity. It's 'The Big Idea' that makes sense of everything.
In Summary: Aggregation Theory describes how certain platforms companies (eg. Facebook, Google, Uber) have been able to dominate their industries in such a systematic way. It also explains why their valuations are so stratospheric. In the eyes of investors, they are playing a zero sum game.
According to Ben, Aggregators have the following characteristics:
- they have a direct relationship with their users
- the goods they 'sell' are digital and thus have zero marginal costs
- the goods they 'sell' are delivered via the Internet, meaning zero distribution costs
- their transactions are handled automatically via credit cards or PayPal etc
Aggregators enjoy winner-take-all effects. As the value of an aggregator to its users is continually increasing, it gets harder and harder for competitors to take away users or win new ones.
There are different types of aggregators (Netflix is different from Uber which is different Facebook) and there are even Super Aggregators. But their potential is the same: as customer acquisition costs decrease over time, marginal customers are attracted to the platform by virtue of the increasing number of suppliers.
As advances in AI make prediction cheaper, Amazon’s strategy could change from shopping-then-shipping to shipping-then-shopping, says HBR's Ajay Agrawal.
In Summary: AI is fundamentally a prediction technology. So, what happens to a company like Amazon as their data scientists, engineers, and machine learning experts dial up the accuracy on the prediction machine? The answer is that prediction becomes so accurate that it's more profitable for them to ship you goods it predicts you will want rather than wait for you to order them.
If implemented today, the cost of collecting and handling returned items would outweigh the increase in revenue. But, better predictions will attract more shoppers, who will generate more data to train the AI, which lead to even better predictions creating a virtuous circle of increased revenue and lower returns.
Competitive analysis is a waste of time if you don’t understand how it impacts your strategy, says Philosophie's Chris Butler.
In Summary: Competitive analysis is essentially how your strategy works in comparison to your competitor’s strategy. There are real benefits to understanding how your competitors perceive the world and how your (potential) customers feel about them.
Competitive analysis can help frame your own product context, discover other problems your customers have and even bond the team together against a common foe.
But making too many choices because of your competitors, rather than for the outcomes your customers need, will put you at their mercy. No matter what customers think of your current product, you're fighting against their perception of your competition, not their product's features.
The deeper the skillset required to perform a service, the less likely the traditional Uber model will work, says PocketSuite's Sam Madden.
In Summary: Before you start building the next 'Uber for X' product, think through some basic variables.
On the consumer side, think through what type of service industry you are trying to disrupt and where that industry falls on the skill vs. urgency graph. This will help maximise client convenience. On the pro side, think through customer conversion vs. retention needs, and optimise for the highest and most economical utilisation rates.
Your product will need to take into account customisation services or projects needed, an understanding of whether the pro is right for the job and the flexibility of communicating the request.
Maximising value means providing maximum convenience for the client and maximum utilisation rates for the pro.
Anyone who watched Jian Yang's app label all food as either 'Hotdog' or 'Not Hotdog' in Silicon Valley may not realise how accurate this scene really is, says James Somers for MIT.
It's the new normal for every startup and tech company to talk casually about how they're 'using AI' to accelerate their product's value. In most cases, these claims overlook 2 fundamental realities. The first is the sheer size of the data sets necessary for machine learning to be applicable, the second is how limited the technology still is.
To get a person to recognise a hot dog, you show her a hot dog. To get a deep-learning system to recognise a hot dog, you might have to feed it 40 million pictures of hot dogs.
Local challenges for Product Managers can be very different from those in developed countries, says Google's Alyssa Maharani
In Summary: For much of the world's population, the Internet is still slow and unstable. Ensuring your product requires minimal data to operate is key to ensure adoption. It's also essential to incorporate some form of 'offline mode' to ensure that customers can continue using your product when there is no Internet coverage at all.
English continues to be the language of social status so products can still be designed in English. But Product Managers should consider adding symbols and icons to help those with less understanding of the language.
Many people buy secondhand phones in emerging markets so products must be optimised for older makes, slower processors, and lower memory space. To test a new release, go to known 'blindspots' in the country, buy some secondhand phones and continuously gather feedback.