This is part of a series of posts on Product Thinking:
Product thinking is about being able to think about products through many different lenses. There are different processes, disciplines, frameworks, formulas, and ideologies that all have valuable lessons for someone who wants to make their product successful for their business and users. This is the first in a series of posts where I share examples of these ideas.
Awareness at scale is ultimately about being in touch with reality by being able to perceive system wide dynamics in a product or in a market. Doing this requires an understanding of common types of dynamics and an intuition for recognizing and distinguishing them. Another important component, which will be a theme across other posts as well, is about being able to recognize and free yourself from your own experience and biases.
Innovation, Optimization, and Novelty
There’s so much value in being able to reason about a market and your own product or ideas in its context. I’d like to introduce a framework through which you can organize this thinking. In these definitions, success is measured by things that make your business successful (usage, revenue, retention, etc). The reason why I think this is valuable is because it ideally helps you reason about what to do with your own ideas and how to judge those of others in determining what has merit, potential, and is ultimately worthy investment.
Innovation is when you create something that is fundamentally different that solves an unsolved problem. Optimization is when you take something that is solved but you solve it slightly but definitively better. Novelty is insidious and commonly misperceived as innovation or optimization. Although I think there is fundamental value in novelty as it is something that people will infinitely seek, novelty in this framework is temporary and lacks any real impact compared to innovation or even optimization. I’m referring to things like very local optimizations and novelty effects which can lead to misunderstanding success or failure and a further waste of resources. If you are rigorous with yourself about bucketing your ideas, your product, and the market into these categories you get a clearer picture of where things stand and what’s worth investing in. Below is an illustration of the innovation → optimization S curve. Innovation causes the highest rate of change but optimization still moves us forward. Eventually in a period of optimization, a new innovation occurs rendering more optimization ineffective.
Mobile phone hardware and software is a pretty digestible example of this. I feel fairly confident in saying we are at the top of the S curve for hardware and mobile software interaction design. To focus on software, we have as an industry coalesced upon a fairly standardized solution space. We know, for the most part, what the best interactions are for navigation in a mobile space and users have developed expectations and inclinations that are hard to combat. We know how to make things look interactive, how to establish modality, how to give feedback, etc. We are of course still making some optimizations like incorporating haptic feedback or moving more UI towards the bottom of the screen near our users’ thumbs, but we are mostly at parity compared to 10 years ago.
In general once an innovation happens, there is definitely room for optimization but the market tends to not have the appetite for much of it. Often what seems to happen is that people still make products anyways trying have a share of the market while only presenting novelty and perhaps some optimization. I think this represents both an over estimation of the liquidity of the market and also a misunderstanding of effects of different optimizations. I think of Path as a product that suffered this fate. They took hundreds of social systems that had been working for Facebook and “optimized” a handful of them with the hypothesis that this was enough of an innovation to have a share of the market. They also over valued some novel aspects of their product like their interaction design. This ended up not working for them because there ended up not being a compelling enough innovation for users to overcome the switching costs of leaving a place with strong network effects (see below). Once there is an incumbent and you only have optimizations to offer, it’s really hard to grow your user base by stealing users from the incumbent, so being able to see the difference between optimizations and innovations is a powerful skill.
Another thing to consider is that the innovation can be non user facing. To use an analogy, let’s say company A produces a physical good and company B produces the same good but it’s a different color. From a consumer’s perspective that appears as just a novel difference. However let’s say that company B found a way to make the same product 20% cheaper than company A. Their innovation is in their production process and not their product. In software this translates to process, culture, or organization design.
One trap that many people fall into is thinking that their individual experience is representative of the average user. This not only ignores the effects of scale but also things like the realities of machine learning (or any dynamically) generated experiences. A common thing I see is people making assumptions without acknowledging that they are in a particular cohort. Someone might say: “My twitter feed is all about politics, so that means Twitter as a whole is all about politics.” Not only is it possible that people like youare having a different experience, it’s possible that the majority of users are having a different experience. Regardless of wether it is true or not, it should be something you are capable of considering. In addition to misunderstanding the type of experience a user is having, it can be costly to make an assumption about the average quality of experience. Often users that use a product with different frequency (new users vs. power users) are experiencing quality in different ways. I think the best way to understand things at scale is by looking at data to get a top down view of different behaviors. You as a product person might find that some feature makes sense to you and everyone you know, but if you look at the data, the average user does not understand it because they aren’t even using it. A good example would be the hamburger menu interaction pattern (https://www.quora.com/What-are-t...). This is something that made sense as a design pattern and worked for people who have spent a reasonable amount of time looking at UI elements and could deduce what that icon meant, but when you look at the data, the average user did not even see it.
Another way to think about scale is by looking for power laws.
The power law can be used to describe a phenomenon where a small number of items is clustered at the top of a distribution (or at the bottom), taking up 95% of the resources. In other words, it implies a small amount of occurrences is common, while larger occurrences are rare. For example, where the distribution of income is concerned, there are very few billionaires; the bulk of the population holds very modest nest eggs.
Abhinav wrote a great answer explaining this in another way: https://www.quora.com/What-is-the-intuitive-explanation-of-Power-Law/answer/Abhinav-Sharma
Examples of power laws are everywhere in software. Some more macro examples would be the distribution of traffic across all websites and the distribution of apps that actually get downloads. Here’s a distribution of AOL users across different sites in 1997.
One can observe that a few sites get upward of 2000 visitors, whereas most sites got only a few visits (70,000 sites received only a single visit). The distribution is so extreme that if the full range was shown on the axes, the curve would be a perfect L shape. Figure 1b below shows the same plot, but on a log-log scale the same distribution shows itself to be linear. This is the characteristic signature of a power-law. http://www.labs.hp.com/research/...
This is both due to a lack of long tail of value (generally a random restaurant website is a worse search result than that restaurant’s yelp page), and some conservatism / exploitation in the ranking of the systems we use to discover web sites or apps (Google search, App Store).
An example of an in product power law is the general behavior on internet products around content production. I’ll give three variations of the same idea:
- The 1% rule is that only 1% of the users of a website actively create new content, while the other 99% of the participants only lurk.
- The 90–9–1 rule states that in a collaborative website like Wikipedia, 90% of the participants of a community only view content, 9% of the participants edit content, and 1% of the participants actively create new content.
- The 80/20 rule, known as the Pareto principle, states that 20 percent of a group will produce 80 percent of the activity, however the activity may be defined.
Recognizing power laws can be a pivotal in anyone’s thinking about products because it helps you understand that not every user is the same, not every entity in a system is a same, and therefore the decisions and investments you make can be higher leverage. If 20% of your users are generating 80% of the value of your product, then you might prioritize work or make tradeoffs to make their contributions even more impactful.
A network effect (also called network externality or demand-side economies of scale) is the positive effect described in economics and business that an additional user of a good or service has on the value of that product to others. When a network effect is present, the value of a product or service increases according to the number of others using it
Simply put, something with a network effect gets better as more people use it. A commonly used example is the telephone. The value of the innovation and the quality of the individual user experience increases as more people have telephones. Typically any network or marketplace benefits from growth in this way. We can use Craigslist as a software example of this. Craigslist is a terrible product with strong network effects. In a sense, the minimalism of Craigslist’s product offering means that the entire thing is mostly just a network effect. Craigslist doesn’t do anything as a product so therefore it’s hard to do anything better than Craigslist. Typically it’s very hard to destabilize the network effects in a product in order to compete with it. In the case of Craigslist, innovation would have to come from a new place and that product would have to build their network up from a strong premise of innovation. From the perspective of a user, the switching costs of porting over your network to the new network are never really worth it and may be impossible. I’m sure someone could make a really nice competitor to Craigslist with a bunch of optimizations but everyone in the network would have to collectively agree to switch. A significant innovation or new paradigm needs to be introduced in order to compete with a product that has strong network effects. Telephones have decreased in importance as new completely different types of communication have become good. Snapchat became successful because they created a network with very different rules that appealed to a part of the population that Facebook was no longer serving. If you want to go deep on Network Effects, I’d recommend the Network Effects chapter of this book.