Category Archives: Insights

About Me

I’ve been an internet entrepreneur since 1997. I co-founded the Apple Music Group in 1995, worked at N2K (one of the first online music companies), co-founded MyPlay (pioneer of the digital music locker), and was COO/CEO of eMusic for five years. Now, I’m a Partner at Venrock in NYC focused on investing and helping build early stage internet and digital media companies. My areas of investment focus include ad tech, social/mobile media, consumer services, web services, consumer products, AI, SaaS, crypto currencies and anything else hugely exciting and disruptive. I am on the board of Dstillery, Smartling, YouNow, Amino Apps and Misty Robotics, and I led our investment in Dollar Shave Club (acquired by Unilever for $1B), Klout (acquired by Lithium Technologies). I oversaw our investments in Crunchyroll (acquired by The Chernin Group) and Nest (acquired by Google).

Follow my tweets at @pakman.

On this blog, I aim to cover various topics related to technology startups, entrepreneurship and the changes in the media industry. I like to chronicle the various machinations of incumbent companies as they attempt to adapt to the massive disruption forced upon them by technology. I very much enjoy the process of meeting entrepreneurs and discussing the exciting challenges of building disruptive companies. I hope you will contact me.

On the personal side, I love being a husband and father of three. As you might guess, I have deep passion for music, and actively drum, record, write, play gigs, and DJ. Tennis, photography, biking, fitness, food and travel probably round out the picture. If you are a entrepreneur building a great new startup, I’d love to hear from you. While you can certainly email me directly (dp at venrock dot com), the best way to engage is with a warm introduction through a mutual connection. We must both know some of the same people…

Follow @pakman


As a Partner at Venrock in New York, David focuses on early stage venture investing in consumer and enterprise internet companies. His active investments include Dstillery, Smartling, YouNow, Amino Apps and Misty Robotics. He led both the Series A and Series B rounds in Dollar Shave Club and sat on the board until that company was acquired by Unilever for $1B in July of 2016. He is a former board member of Crunchyroll (acquired by The Chernin Group), led Venrock’s investment in Klout (acquired by Lithium Technologies) and oversaw Venrock’s investment in Nest (acquired by Google). In April 2017, he was named the 55th top venture capitalist by CB Insights.

Previously, Pakman was the CEO of eMusic, the world’s leading digital retailer of independent music, second only to iTunes in number of downloads sold. After buying eMusic from Vivendi Universal, in the three years that Pakman ran it, he grew the business by more than 850%, from $7M in revenues to more than $68M. Pakman transformed the business from an obscure also-ran with 50,000 subscribers to the second largest digital music retailer in the world, with more than 400,000 subscribers and more than 12% market share.

Prior to joining eMusic, Pakman was Co-Founder and President of Business Development and Public Policy at Myplay, Inc., the company he co-founded in 1999 in Redwood City, CA that introduced the “digital music locker” and pioneered the locker category. In 2001, Myplay was sold to Bertelsmann’s ecommerce Group. Before Myplay, he was Vice President at N2K Entertainment, which created the first digital music download service. He also was the co-creator of Apple’s Music Group and a product manager at Apple.

Pakman is a graduate of and a former member of the Board of Overseers at University of Pennsylvania’s School of Engineering and Applied Science with a degree in Computer Science Engineering and is an avid musician and songwriter. He testified in front of Congress regarding internet music streaming in 2012, 2016 and 2017. He serves on the board of POTS and is a Trustee at St. Luke’s School.

Since 1969, Venrock has developed into one of this country’s premier venture capital firms with offices in Palo Alto, New York, Cambridge and Israel. Venrock continues the eight-decade Rockefeller tradition of funding entrepreneurs and establishing successful, enduring companies. We have invested $2.6 billion in over 450 companies over the past 42 years, resulting in 128 IPOs and 137 M&A exits. Venrock’s returns place it among the top tier venture capital firms that have achieved consistently superior performance. Venrock currently oversees 121 portfolio investments across the information technology, healthcare, and energy sectors and has been behind pioneers including Adify, Adnexus Therapeutics, Apple Computer, Athenahealth, Centocor, Check Point Software, DoubleClick, Gilead Sciences, Idec Pharmaceuticals, Illumina, Intel, Ironwood, Millennium Pharmaceuticals, Sirna Therapeutics, StrataCom, and Vontu. For more information, please visit Venrock’s website at

About Me was originally published in on Medium, where people are continuing the conversation by highlighting and responding to this story.


The Inevitable Evolution of Online Sharing — Live Video Conversations

It’s been an incredibly exciting few weeks as the world comes to hear more about the latest category of social sharing — live mobile video. (Disclosure: we have been bullish on this space since our investment last year in YouNow.) Given that we have had the pleasure of observing this phenomenon for a bit, I thought I would share a few thoughts.

It’s Not Broadcasting, it’s a conversation
Many people are calling these live video feeds “broadcasts” which presume they are one-to-many and are one-way. The internet has taught us that all media must be participatory now. We all have an expectation that we are not just observers, but that our voices as viewers must be part of the content itself. From likes to comments, the web is built on the “post-respond” engagement model. For any of these apps to be successful, they must engender engagement. And to be engaging, the viewer must be a participant in some way. In YouNow “broadcasts”, the chatting audience is as much a part of the content as the “broadcaster”. This is what leads to successful long-term engagement.
utility vs. platform
Meerkat launched as a livestreaming utility built atop the Twitter network. Like and Twitpic before it, Meerkat itself is not a platform. One does not browse Meerkat to find content, one waits for announcements in the Twitter stream. If they are useful, utilities for social networks very quickly get absorbed into the platform as a core function, leaving no room for third parties. We saw that with link shortening and image hosting, and now we are seeing it with Periscope.

However, much like YouNow, Periscope is more of a network. It uses your existing Twitter graph to build your Periscope graph, but ultimately users will prune and grow their Periscope graph to look very differently than their Twitter follow graph. The Periscope app includes some (limited) forms of content browsing. I expect, given what happened to Meerkat, and with the very talented Josh Ellman as a board member, they will quickly move in the platform direction. Platforms are much more valuable and more sustainable than utilities.

native media format
All successful social networks have a native content form to them in which users become expert. There is a form to a great Facebook post (baby and party pics), a great Tweet (witty observation and link to interesting news), a great Instagram pic (beach and sky pics with awesome filters), a Vine vid (successful loop), etc. So too with live video streams. We can see it on YouNow, as users become expert at creating engaging performances and successfully interact with (and involve) their audiences. I expect we will see the same with Periscope streams (just not yet. Give it time.)

I really think the promise of these products is in creating more conversations between creator and consumer, rather than being millions of new cameras for livecasting the world. It’s tempting to think of this in terms of crowd-sourced news-gathering, which is a compelling use case, for sure. But the web has taught us that social media must be interactive to be successful. Sure, it will be supercool to watch Mario Batalli cook in his kitchen. But that’s the TV model. Unless he personally interacts with his audience (and is really good at it), I don’t think it is web native enough to work. So I suspect the winning formats for all of these products are the ones which are most participatory.

The Inevitable Evolution of Online Sharing — Live Video Conversations was originally published in on Medium, where people are continuing the conversation by highlighting and responding to this story.


Here is what Jay-Z should have launched with Tidal

“The transition from CDs to digital hasn’t worked well for the music industry. Sales are down and too many people listen to music without paying for it. We think that is because digital music is too expensive for the value it delivers. For too long, music has been both too expensive for fans and doesn’t produce enough money for artists. We wanted to completely change the game. So we did.

We, the sixteen superstar artists on this stage, used our incredible power and leverage over the music industry to demand completely new economics from the labels and publishers. So today, we are launching Tidal5. For $5 a month, you get music streaming of 20 million songs to any device. And to help artists, we are introducing the same economics as the iTunes and Google Play stores into music. 70% of all Tidal5 revenue will go to the artists and 30% will be split among the operations of the service and to the labels and publishers.

We wanted to find a way to attract more buyers into digital music and we knew the only way to do that was to get prices much lower. That’s a gift to our fans. But we also needed to get way more money into the hands of the artists. And we did that too. We used our power for good. And we hope you enjoy it.


Here is what Jay-Z should have launched with Tidal was originally published in on Medium, where people are continuing the conversation by highlighting and responding to this story.


Churn is the single metric that determines the success of your subscription service.

A model to help you evaluate consumer subscription businesses (download here)

Not every direct-to-consumer business should be a subscription service. Many try to be, and for good reason—subscription businesses often have more predictable revenue than other businesses, they allow you to acquire a customer once and then have a long-term paying relationship with them, and they can quickly create enormous enterprise value.

However, in evaluating whether this model is right for a particular business, entrepreneurs need to have an intimate understanding of the immutable force governing their success—churn. Having run and invested in quite a few consumer subscription businesses over the years (eMusic and Dollar Shave Club being the most well-known), I have put together a view on churn that I am happy to share.

As I have discussed before, one of the clear indicators DSC could be huge was their low churn rate. This was apparent even in the very early months of the service’s life. On the flip side, many of the meal kit services we looked at showed very high churn. As we will discuss below, this is the single most important factor predicting a service’s future success.

Are Subscriptions Right For You?

First, not every business or service should be sold using a subscription model. Customers tell us if they like buying our product via subscription by how many of them stay in the service versus how many defect. High churn businesses indicate customers would likely prefer another model, like a la carte sales.

For subscription services to have to become large venture-scale returns, two things must be true: (1) the churn rate must be low and, (2) the market size must be large. There are plenty of niche subscription businesses where customer lifetimes are long, but the population interested in them is small. Those businesses are typically hard to reach venture scale and be worth, say, half a billion dollars or more. Some examples of these include good old-fashioned magazines and wine-of-the-month clubs.

What is Churn?

Many subscription companies measure and report churn differently. To properly calculate churn, let’s first establish some definitions. Let’s assume the billing period for a service is monthly, but you can replace “monthly” with whatever your billing period is (ie, quarterly, etc.)

  • A = active cancellations in month (customer took action to cancel)
  • P = passive cancellations in month (credit card declines, non-renewals)
  • W = winback re-activations in month (previous subscribers who reactivated)
  • T = Total paid subscribers at beginning of month

The purest and most complete way to calculate churn is to follow the formula below:

Average Net Monthly Churn = ((A+P)-W)/T

This gives you the rate at which customers are leaving your service as a percentage of the total customers in your service. You should calculate this each month as a total across all cohorts.

What is a Good Churn Rate?

Generally speaking, for consumer subscriptions services to become a large business, I look for average net monthly churn rates below 5%. The truth is, very few achieve this. Almost all of the consumer subscription businesses I have evaluated as potential investments have churn rates higher than this, and in some cases, much higher. The biggest subscription winners are below 5% per month, and usually well below 5%. (For example, Netflix is below 1% per month, Dollar Shave Club is very low, ISPs and Pay TV are also low with Dish at 1.5% per quarter, Verizon Wireless and AT&T are about 1.5% per quarter)

Why is a monthly churn rate much higher than 5% a problem? Well, think about your service as a bucket. Each month, you do some combination of paid marketing and organic customer acquisition to grow your business. Those new subscribers are placed in the top of your bucket. Churn is the hole in the bottom of the bucket. No matter how many new subs you put in the top, you continue to lose customers as they leak out of the bottom. Just to stay flat, you have to replace your churned customers. If a lot leak out, you have to spend/market more to fill the top.

The law of percentages has a nasty effect here as well—the bigger you get, the more customers you lose (and have to replace) every month. The larger your business gets, the bigger the holes in the bucket. A 10% monthly churn rate, when you only have 50,000 subs, means you only lose (and have to replace) 5000 subs a month. But 10% churn when you have 500,000 subs, means 50,000 new subs are needed per month, just to tread water.

A churn rate of 10% a month effectively means you lose (the equivalent) of all of your customers every 10 months. Those businesses are not sustainable. This is a hard point for some subscription novices to internalize—if you lose all of your customers each year, you do not have a good business. You might have a product or service that some customers want, but you are selling it to them in the wrong model. Your customers are telling you they don’t want to buy every month.

One common reaction to this point of view is, “Hey, if some of my customers stay and love the service, who cares if a bunch leave? I will just focus my marketing on more of the people who stay!” This is a fallacy. The basic methods of customer acquisition are all based on targeting to your best prospective customers. Churn is a reaction to your value proposition after people have tried your service. If it is high, you have a problem with the service, not with your targeting.

Period Churn and Cohorts

Many operators of subscription services will note that early customers churn at higher rates than customers who have been with you a while. Period churn is the amount of churn that happens in the same period for each batch of new customers, or cohorts. For example, first month churn is the average churn that occurs in the first month for all new customers. (Here is a good primer on cohorts.)

First period churn is always highest, period two is lower, period three usually lower still, and then things tend to steady-state for the remaining periods. The graph below illustrates a typical churn curve, expressing period churn, or the churn of customers in each of the monthly cohorts.

The math behind calculating period churn is as follows:

A = Number of new subs acquired in a month x

B = Number of subs remaining from A in month x+1

C = Number of subs remaining from A in month x+2 (you can include winbacks in C)

First Period churn = (A-B)/A

Second Period churn = (B-C)/B


Average Period Churn = the average of all similar Period Churn results (ie, all month 1 period churn results)

Cohorts are a great way to study how customers with similar tenures behave and where best to make improvements in your service. Obviously, reducing first period churn will have the biggest impact on your business. Why are so many people leaving in the first month and can we address their concerns?

In my experience, almost all subscription businesses follow a similar curve to the one above, albeit with different values along the periods. Focusing on early-period churn obviously pays the biggest dividends. Important to note—the fact that your service tends to steady-state in the out months is not unique to any one service.

How to Analyze Churn

I created a standard churn analysis worksheet that I use to analyze churn in subscription services. This template allows us to use standard definitions of churn rates and make sure we are comparing apples to apples when thinking through investments in new services. You can download a copy here. Just fill in all the colored cells (and you can extend the sheet to include more than seven periods). This worksheet calculates a number of useful pieces of data:

  • Average net monthly churn — this is the global expression of churn for your service.
  • Average churn for each billing period—this shows you how many users leave the service each billing period, on average
  • Churn for each billing period across your tracked cohorts—this helps you see whether newer users are exhibiting better or worse retention than older users.

Reducing Churn

There is a laundry list of optimizations subscription services implement to improve retention, and collectively these have a positive effect—re-billing insufficient fund accounts on the 15th and 30th of the month, allowing members to pause their service, allowing members to reduce the frequency of delivered goods, winback campaigns to churned members, etc. But a key observation about businesses which show high churn in their early stages is that the churn will not meaningfully reduce. That is, you can expect your optimization efforts to maybe move churn 1% — 2% (absolute points) or so, but they will never halve churn or reduce it dramatically beyond this. This is because churn is a statement from your customers about how habitually valuable your service is to them. In my experience, the only way to meaningfully reduce churn in high-churn businesses is to essentially redesign the service and the value proposition. For clarity, I have looked at close to a hundred subscription services and have never seen churn improve from, say, 12% a month to 7% a month without fundamentally changing the service…but I have seen it move from 12% to 10% or thereabouts through optimization.

In addition, churn can actually get worse over time as you scale up your customer acquisition efforts. Sometimes your enthusiastic early adopters were your very best customers and as you reach into more and more segments to find new users, the mass market users can turn out to be less loyal, hence churn can worsen.

High Churn Eventually Bites You

There are plenty of examples of high early growth combined with high churn rates. Investors got excited about the emerging high-growth meal kit market, probably because paid marketing was working very well. That is, in the early days, the services could buy customers for a CAC (customer acquisition cost) that is less than the margin LTV (lifetime value — the sum of all gross margin dollars over the lifetime of the customer) of every customer. This is good math, right? But when the churn is high, the company must spend more every month to refill the leaky bucket and eventually the business hits a wall. CACs continue to rise, total spending starts to get out of hand, and investors get scared. The business is forced to reduce marketing spending to reduce cash burn, starts to shrink, and then things get ugly.

A recent example of this is Blue Apron. Their average monthly churn appears to be quite high…perhaps as high as 12% per month. With one million total paid subscribers, they lose about 120,000 subs per month. The cost to replace these lost subs, just to stay flat and not grow, at an estimated CAC of $147, is $17.6 million per month. More recent CACs are $460, so they would spend more than $55 million per month just to tread water. You can see how this becomes unmanageable.

The Takeaway

High churn businesses (roughly greater than 5% per month) are not great businesses. As they grow, more of their marketing must be devoted to replacing churned customers. They eventually hit a point where shareholders lose patience with the amount of marketing spending necessary to fund growth. When marketing spending is reduced, if the churned customers are not replaced each month, the company can begin to shrink rapidly. This treadmill gets very hard to maintain and ultimately is the undoing of many high-churn subscription businesses. Things may look exciting in the beginning, but the leaky bucket always catches up with you eventually.

The recipe for success in subscription businesses is a combination of low churn and large applicable markets. If your churn is high, switch to a different model or redesign the offering to find a lower churn proposition.

Churn is the single metric that determines the success of your subscription service. was originally published in on Medium, where people are continuing the conversation by highlighting and responding to this story.


Learning Effects, Network Effects and Runaway Leaders


There’s a new economic force at work in the machine learning revolution that is capable of generating increasing returns to scale, much as network effects did in the internet revolution.

This force is automated learning, and its business impact comes in the form of learning effects: the more a product learns, the more valuable it becomes.

Learning effects have the potential to generate enormous economic value, as network effects do, if companies are able to close this loop and make it self-reinforcing: that is, if their products learn more because they have become more valuable.

This happens when more valuable products attract more users or customers, who provide more and richer data of the kind that enables machine learning models to make these products more valuable still, which attracts more users or customers still, and so on, creating a self-perpetuating cycle.


Just as network effects determined who the biggest winners of the internet revolution were, learning effects will determine who the biggest winners of the machine learning revolution will be.

Because they enable increasing returns to scale, they will similarly give rise to a set of companies that become runaway leaders – that are capable of pulling away from their competitors and continuing to increase their leads over time.

Offline Origins
Like network effects, learning effects have always existed in the offline world but have become supercharged in the digital world. In the offline world, learning effects are transmitted through humans: as people learn how a product can become more valuable, they modify it accordingly. Human learning, however, is artisanal, and artisanal learning only scales so quickly.

What’s new and different in the machine learning era is that certain kinds of learning have become automated. Software can learn by itself with exposure to new data and become more valuable in the process. This is a big deal economically. It involves the unlocking of new source of economic value that was previously inaccessible.

A Vast New Power
Learning effects have taken off most significantly at large internet platforms given the immense amount of data they control and their aggressive investment in machine learning to accelerate product innovation: Google in search, ads, photos, translate and Waze; Facebook in search, ads and newsfeed; and Amazon in search, ads, product recommendations and Alexa, to name just a subset for each. These companies recognize that machine learning has granted them a vast new power, and they are eager to take maximum advantage of it.

Perhaps the best pure-play example of the power of learning effects is Tesla, which began as an electric car company but was able to deploy machine learning to extraordinary effect across its fleet to become the category leader in autonomous driving. Tesla’s autonomous driving capabilities make its cars more valuable, which attracts more customers and data, which enables it to improve these capabilities further and attract even more customers, and so on. As a result of its learning effects, Tesla’s rate of innovation and value creation in the autonomous driving area have dwarfed what its competitors have been capable of.

Engineered Growth
Network effects and learning effects generate growth in different ways. Network effects tend to generate growth organically through a kind of gravitational accretion, as individual consumers and businesses pursuing their own self-interest decide to join the largest and most valuable networks, making them larger and more valuable still.

Learning effects similarly benefit from consumers and businesses pursuing their own self-interest to purchase the best products, but they are less the result of gravitational accretion than of finely tuned technology and product development efforts that require constant intervention and recalibration in order to tie together data, intelligence, product innovation and user/customer growth.

As a result, even though learning effects are partially the product of automated learning, they are by no means automatic. The data generated from new customers must be of the right kind and of sufficient volume to enable new learning. This learning must be optimized effectively enough to create new product value. And this value must be strong enough and productized well enough to attract more customers. Any break in this chain means there is no self-reinforcing cycle and hence no learning effects.

Runaway Leaders
Perhaps the most interesting question about learning effects is what are the conditions that make them strong enough to create runaway leaders, as these are the companies that tend to create the vast bulk of enterprise value in the technology startup world.

Learning effects don’t always produce runaway leaders. Just because one company has a head start in learning doesn’t mean other companies can’t acquire more or better data or learn more efficiently from similar data to catch up with them and eventually bypass them. It’s an interesting question today, for example, if Tesla is pulling away from the pack in autonomous driving, or if others will catch up in the years ahead.

In order for learning effects to produce runaway leaders, a company must secure a definitive advantage over its competitors in one of the component areas of learning effects – data, intelligence, product innovation or user/customer growth – and leverage this into advantages in the others, such that the company can acquire data, learn, innovate and grow not only more rapidly than its competitors do, but more rapidly than they can.

As with learning effects generally, there is nothing automatic about tying these advantages together. It requires excellent execution.

Typically a company is able to jumpstart this cycle by developing a significant data advantage over its competitors. It then must translate this data advantage into an intelligence advantage as measured by the capabilities of its machine learning models, which requires that its models be as or more efficient than those of its competitors. This intelligence advantage must then tie to a product innovation advantage that is directly correlated with a user or customer acquisition advantage and ultimately with an advantage in the size of its user or customer base. Enough customers have to want to buy Teslas, in other words, because of their autonomous driving capabilities vs. because it’s an electric or cool-looking car, as that doesn’t create a strong enough self-reinforcing cycle. Finally, this user or customer base advantage must enhance the company’s data advantage in the right way to generate additional learning.


Generally the narrower the scope of a product and the greater the degree to which machine learning drives its value, the easier it is to tie these advantages together to create a runaway leader.

Wherever it is possible to tie these advantages together, there will likely be ferocious competition, as with network effects, for startups to get an initial head start in competing for scale to achieve escape velocity and become runaway leaders given the huge premium on winning. The early bird that capitalizes on its head start generally gets all the worms. Other birds need to bootstrap alternative advantages in the form of more efficient learning engines or access to large and differentiated datasets in order to have a chance.

Learning Curves: Long, Steep and Perpetual
In order for runaway leaders to be able to maintain their leads over time, there’s an important additional requirement, which is that the learning curves for their products must be long enough and steep enough to enable them to provide increasing product value for an extended period. If the learning curves for their products are short or top off quickly, early leaders will max out on them while they still have viable competitors, and these competitors will be able to catch up. If the learning curves are long and steep, on the other hand, these companies will have sufficient runway to break away from their competitors and maintain their leads over time.

Certain products – particularly those built on highly dynamic datasets – may have perpetual learning curves such that in a rapidly changing world, they can always be meaningfully improved. It’s around these kinds of products that the most valuable runaway leaders will likely develop. Potential examples include search, semantic engines, adaptive autonomous systems and applications requiring a comprehensive real-time understanding of the world.

The Interaction of Learning Effects and Network Effects
Network effects almost always create the opportunity for learning effects, as they involve the generation of ever more data in the form of new network members and interactions. Companies must invest in machine learning to create these learning effects, and they may or may not be successful. They may fail to generate meaningful learning, or they may generate meaningful learning but not learning effects if this learning does not result in more valuable products that lead to the continual acquisition of new data for additional learning.

Conversely, learning effects can create network effects. Tesla, for example, did not benefit from network effects when it was just an electric car company and was not yet focused on autonomous driving. However, once the company outfitted its cars with information sensors to develop autonomous driving capabilities through machine learning, it suddenly began to benefit from network effects: each Tesla became more valuable the larger the fleet became.

Importantly, however, when learning effects create network effects, these network effects do not exist independently of them. They are in effect an expression of the learning effects: learning just happens to take place through a network. If Tesla turned off its machine learning, its network effects would cease to exist.

The reverse, however, is not true. Network effects can give rise to learning effects that can exist independently of them. Facebook’s core network effect of people wanting to be part of the same social network that their friends are, for example, generates lots of new data that machine learning models can learn from. One area where Facebook has invested significantly in machine learning and succeeded in generating learning effects is improving the relevance of its newsfeed. Newsfeed relevance is a different kind of value than the core value around which the company’s network effects are based, although the two clearly reinforce each other. If Facebook stopped growing its user baser, it could continue to generate increasing value by improving the relevance of its newsfeed through these learning effects.

Since network effects and learning effects are both functions of customer value, whenever they exist side by side in a product, they always reinforce each other, as each makes the product more valuable in a way that attracts more customers and data.

The most formidable kinds of runaway leaders that tend most strongly toward natural monopoly – Facebook and Google are excellent examples – are those that benefit from network effects and learning effects working in tandem, as their mutual reinforcement means these companies run away from the pack much faster and are generally impossible to catch, provided they also benefit from perpetual learning curves.

Startups vs. Incumbents
Incumbent internet platforms have unsurprisingly been the big winners of the machine learning revolution to date because of their vast data assets and their significant investment in this new technology. Their early dominance has led skeptics to wonder if machine learning is a game that startups can win at all given their relative data disadvantages.

There are huge new datasets and data-rich applications created every day, however, in domains where these and other platforms have little or no presence, which provide an abundance of new opportunities for startups.

In addition, there are many large datasets sitting in organizations that startups are best suited to access because they are better able to provide these organizations with innovative applications to take advantage of them.

And although startups make lack the early edge in data, they always have the advantages of focus and adaptability. Where I believe these advantages will make the biggest difference in machine learning is that machine learning applications are engines, and startups have the ability to build and tune these engines most precisely to maximize learning effects. They have the ability not only to maximize the amount of learning and hence value they create from new data, but to complete this loop and maximize the amount of data in the form of new customers they create from new learning.

Only by constantly tightening and amplifying these loops can companies grow rapidly from learning effects and hope to achieve escape velocity to become runaway leaders. As a general rule, startups tend to be better at this than incumbents.

This article was originally published on Techcrunch


A Robot For Every Home and Office?

Looking out 20 years, do you think we will all have a Rosie-type robot in our homes, doing chores and fitting in as a member of our family? Most of us believe this future is likely. But how do we get to there?

Misty Robotics, from the makers of Sphero, the leaders in connected play products like BB-8, believes they have an answer. They view the path to universal robot adoption similarly to how other major tech products evolved, such as the personal computer. The world is not yet ready to buy a personal Rosie robot on Amazon. Instead, the path to this future is likely composed of many methodical steps.

The first PCs had very few applications. They largely appealed to innovators and developers as a platform on which applications could be built. The Apple ][ ads proclaimed that it shipped with BASIC and your first night you could be “entering your own instructions and watching them work.” It was a highly capable computer with great specs meant to inspire you to program it. Eventually, Apple believed, applications would emerge obviating the need for Apple ][ owners to need to know how to program. When Apple Writer and (later) VisiCalc appeared (both created by third-party developers), the computer took on entirely new purpose and general appeal.

The Misty team believes we are in the same stage of personal robotics. There are more than a million casual and professional roboticists in the world, tinkering and creating robots. But every time someone wants to program a robot to do something, they usually have to start by building the robot itself. There aren’t any fully-capable general purpose consumer robots.

Yet many of the great technology shifts emerge as platforms, on top of which innovators and creators build useful, diverse and ultimately world-changing applications.

Welcome to Misty. It will take time for Misty to become something all consumers want to buy. That is not the early focus here. But we believe the technology is in the right state to enable innovators to create the diverse set of skills and capabilities required for broad market use.

The Misty team is an incredible group of passionate roboticists and consumer product experts who have built and shipped consumer devices that have sold many millions of units. Ian’s Sphero background has taught him and his team plenty about the way consumers buy and engage with connected robots. Tim’s Apple and Nest experiences inform his passion for building high-performing teams who ship elegant and successful must-have products.

Together, this team has a grand vision — one day, we will all have a robot in every home and office, helping us, entertaining us, and joining us as part of our family. We are honored and proud to be along for the ride, along with some fantastic co-investors, like Brad at Foundry. I am also thankful to Paul at Sphero for helping see the promise of this idea and organizing a spin-out to allow it to reach its full potential.

As a partnership, before making an investment, we always ask, “if everything goes right, how big could this be?” In this case, the answer is “huge” and we look forward to trying to help the team build an incredibly meaningful and impactful company.

A Robot For Every Home and Office? was originally published in on Medium, where people are continuing the conversation by highlighting and responding to this story.


2017 Healthcare Prognosis

The state of US healthcare delivery has certainly been top of mind in 2017. With so much potential change, we are frequently finding ourselves in discussions about how policy will change and impact our ecosystem. While never short of opinions, we also have a lot of questions ourselves. Rather than relying on our Magic 8 ball, we decided to crowdsource (a few hundred of the smartest people we know across healthcare) predictions on the future. We covered the impact of the Trump administration on the ACA (Affordable Care Act), how various healthcare IT subsectors will fare going forward, valuation sentiments, amongst other topics.

Our key findings are presented in the Venrock 2017 Healthcare Prognosis. You can read it here.

Running Through Walls: Culture Conscious

When Matthew Prince and Michelle Zatlyn met in business school, they didn’t envision a class project turning into a billion-dollar company, but that’s exactly what happened. Bryan Roberts, partner at Venrock and Cloudflare board member, talks with the co-founders about the early days of company building and how their initial mission statement has remained the same years later, a rarity among Silicon Valley startups. Prince and Zatlyn discuss their measured and thoughtful approach to hiring, and why slower growth helps them keep Cloudflare’s culture strong. They also share their experience with public policy, and a time when they took drastic measures to protect the privacy of a Cloudflare user.

Running Through Walls: Surviving a Slump

Bruce Cozadd was a musician in his early years, but a passion for science and business led him to enter the biopharma industry. Venrock partner Camille Samuels talks to Cozadd, now CEO of Jazz Pharmaceuticals, about his journey to co-founding Jazz and the people who helped him along the way. He shares the joys of starting Jazz with a team that had worked together in the past, but also highlights the downside: that his team lacked diverse prior experiences to rely upon while building a company. When Jazz’s stock price fell to just $0.53 a share, Cozadd persevered and relied on grit and determination to turn the company around. He also shares his wisdom about managing people through all stages of their careers, and reveals what his mentor taught him about treating people well no matter how difficult a situation the company is in.