A Model: Why Speed Reading May Enhance Ideation and Innovation

I took a course on speed reading earlier this year and began “flying” through books (flying, that is, compared to old Frank). Prior to taking this course, I would move across a page slowly, reading each word “aloud” in my head. I would move line by line, carefully contemplating the ideas at the moment as they presented themselves. The trick with speed reading in my experience is that you skip past both of those steps. My good philosopher friend Neal Donaghy would ask me “well, are you really reading then? Are you actually absorbing the information and using it to think critically about the world?” A valid response, and a skepticism I too shared for many years.

For me, there is certainly a trade-off with speed reading. Unlike traditional reading, I don’t retain as much information at the moment while actively reading, but I have found that the increased speed of information consumption does allow me to connect ideas between books more readily. For example, Hal Varien and Carl Shapiro’s Information Rules (1999) discusses the application of durable economic principles to the then, emerging digital information economy and a recurring theme is economic switching costs. Every chapter in some way, shape, or form goes back to this idea. As I read through Vaclav Smil’s Energy, the impact of economic switching costs involved in our decision of how to create energy (nuclear, coal, oil, renewables like solar & wind) echoed loud and clear through my mind. Similar principles, but applied to the physical world rather than the virtual world. It is my belief that the connection of loosely related information creates a ripple effect of ideation, improves long-term memory, fosters innovation, and ultimately enhances a person’s problem-solving capabilities 

Connecting information helps us build a holistic picture of the world around us which in turn gives us new ideas. A relatively simple data model can convey the concept elegantly. To start, a few assumptions…

  • It takes 1 week to speed-read a book, it takes 2 weeks to regular-read a book.
  • The retention when speed-reading a book is 5 (on a scale of 1-10), half the retention of regular-reading which we will call 10 (I refer to this sometimes as the ‘level’).
  • Each week after reading a book, it’s retention – or level – weakens by 1 (you forget over time if it is not reinforced). When the level gets to zero, the idea is no longer available meaning it won’t help you create connections and ideas.
  • Consider the domain of knowledge on a 100×100 grid; when 2 ideas from different books are related defined as within a distance of 20 to each other, they form a connection and enhance their level by 3. Thereafter, they continue to degrade over time.
  • In our simulation, we will consider 100 weeks – just about 2 years.

We can debate and play with these assumptions, but they are a fair starting point. Let’s look at those assumptions visually:

It is my belief that the built-up clusters of related information help us create new and better ideas. Now, instead of showing the first 8 weeks (above) we can see what would happen over the 100 week period with a .gif:

Over time, speed reading provides increased availability of information over regular reading.

While clusters of related information come and go through the period, after 100 weeks you have developed a really nice network in the upper half of the speed reading graph. Generally, we have a trade-off between retention and time. Speed reading reduces time but also reduces retention. In turn, the broader availability of information helps us create more ideas increasing the probability of innovation. While thinking through this model I realized that there was a critical piece missing; powerful anchors.

Let’s create an analogy to competitive racing. Throughout a competitive season, there are these grueling races or workouts that challenge you both physically and mentally. These workouts ‘stay’ with you… especially the mental grit component (for example a very long 20-mile run). Even though they may occur several weeks or months before the race, they are the foundation that holds your training together. Similarly, with reading and ideation, there are likely books that you read that really resonate with you – and the information just lingers. This is likely because you already have some connection with the material. In our model, this would mean these nodes maintain their level… the information stays available for long periods of time. We must account for these books with a hybrid model. I’ll add a new assumption that acknowledges we read 4 books a year that fit this description and create hybrids of the original models.

Considering ‘anchors’, speed reading leads to a beautifully connected network of information.

You get some benefit from the hybrid model with regular reading, but still not to the extent that speed reading gives you. On the other hand, the hybrid approach mixed with speed reading leads to a widely connected network of information. There are three extensions/applications that are worth introducing. 

First, books aren’t the only source of information for us. Quite the contrary. In fact, in 2019 Pew Research reported that 20-30% of American adults hadn’t read a book in the previous year, and, of those that did the median American read 4 books per year (far less than the 25-50 used in the model). It’s probably fair to say that many of us do much more reading of news,  short-form articles, and video/podcasts (this is what I think of as ‘short-form diversified content’). Hypothetically, here’s what that trade-off between retention and time might look like for different forms of media.  

FormLevelTime
Short-Form Article11
Short-Form Diversified Content33
Long-Form Article45
Book (speed reading)57
Book (deep reading)1010

Notice that well-written, long-form articles might play a nice role in your information network construction as they provide almost nearly as much retention as speed reading a book (according to my experience), but take “much less” time to consume. A friend of mine introduced me to longform.org last year and he’s right: when you take 20-30 minutes to read these articles you feel like an expert for a day… and that kind of information stays available for quite a long time. Thanks Chris. 

Second, this model applies to opportunity. You can probably trace back through time and identify events that were unlikely to occur but changed the course of your life. For example, you met the right person in a seemingly random way OR you seized an opportunity that scared the shit out of you OR you connected 3+ ideas to innovate a process. We call these black swan opportunities – very low probability events that create tremendous possibility(1). The more low probability situations you create (more available dots), the more likely some of those situations are to connect to each other and provide you with a way forward. When you can make the right connections at the right time, this leads to those transformative events.

Third, failing fast. As teams all over an organization experiment, they continue to learn (this creates new dots). If experiments take 6-12 months you will never achieve the critical mass of dots you need to form the right connections and learning that drive innovation. It might be unpopular, but I think leaders should encourage rapid experimentation (note I didn’t say ‘reckless’ experimentation). You may not always know what other dots, at what level, are already on the map but the fact here is that the more dots you can get on the page in a relatively shorter timeframe… the more likely you are to create something very profitable. This shows that over time, the company that moves faster will be more profitable even though in the short run may be painful. This same idea exists for venture capitalists – bet on 100; lose small on 98 but win big on 2.

This model isn’t real or supported by actual data, just experience and logic. But, it is a really helpful framework for thinking about how I spend my time, how I allocate my attention, and how I solve problems. More broadly, if you are willing to accept and acknowledge the short-coming of data and data-driven thinking… it can be a powerful tool in your tool belt. Like the disclaimer COVID Act Now was putting on their models’ output, “this model is intended to help make fast decisions, not predict the future.”

(1) Nassim Taleb discussed the idea of Black Swan Events in his books Fooled by Randomness and The Black Swan. Black Swan Events are events with extremely low probabilities, but extraordinary consequences (such as the financial crisis in the US in 2008).


Leave a Reply

Your email address will not be published. Required fields are marked *