The Paul Neto Blog

The Rise of human assisted AI

noreply@blogger.com (Paul Neto) — Thu, 16 Mar 2017 20:23:00 +0000

A three year old child can identify a picture of a dog among pictures of food with more accuracy than a computer (want a few more examples, check here).

Fifty years of computer intelligence development and we are essentially nearing the processing level of a three year old, though in the past five years we've achieved more advancement in this area than we have in the past fifty years. This is all due to the rise of activity in machine learning and what is fundamentally changing how we deal with data and think about insights.

We've moved from creating rule based systems to systems that can learn from data. Machine learning, deep learning, SVM, neural networks, all provide new opportunities. Similarly, the abundance of readily available tools such as R and Python libraries to accomplish processing and analysis not possible just a few short years ago.

For most people, it is a difficult concept to get their mind around though a number of recent developments have started to emerge that highlight the power of this approach. Namely, Tesla and it's efforts around autonomous vehicles. The idea of cars driving themselves is something most individuals can grasp and many may be either fearful or impressed with the concept. Not surprisingly, much of the discussion around artificial intelligence (AI) has been around automation or building autonomous systems that can perform tasks that humans do today. The primary difference with this concept from that of automation in the industrial revolution is back to the core difference of these 'machines' being able to learn as they automate processes.

These learning systems typically work by learning from numerous examples, much like the puppy example in the image above. The more examples, the more attempts, the rate of success increases. This is much how the human brain learns, thus the effort is to build systems that mimic how the human brain works. AI systems can learn either by using techniques to discover patterns in data, or through a process called reinforcement learning. Reinforcement learning is based on the concept on how humans learn, through vast examples of trial and error and establishing a reward system when there is success.

One of the core challenges as a result of these techniques is to have a sufficient number or recurring number of features or examples for the machines to process, find patterns and learn from. Furthermore, understanding the breadth of options to explore becomes increasingly important for reinforcement learning.

Humans have an incredible ability to learn quickly, though are relatively inefficient in being able to transfer knowledge from one entity to another. One of the key benefits of a learning system, is that the intelligence can easily be replicated and applied to other system entities. For example, autonomous cars require the processing of millions of miles of driving, though the true benefit is the accumulated learning across many vehicles and not a single vehicle. Similarly, voice recognition systems such as Siri depend on the collective knowledge to improve it's accuracy.

The true success of AI may indeed depend on the ability to foster the availability of data features and examples for processing. I believe this is where the human assisted element plays an important role. As companies and industries start to adopt these technologies, considerable effort is required to navigate the data elements available for processing and building a reliable data pipeline, even before any machine learning can occur.

One of the richest sources of data for processing is from humans themselves. Increasingly it will become critical for systems to be able to monitor and track human behaviors, their trials, their errors and their successes. How we process data, conduct certain tasks, make our own decisions on what to do next, may provide a rich set of data for learning from.

One area of application where this may become quite fruitful is in digital advertising. Most ad technology of the past decade has been predicated on a process of throwing shit at the wall and seeing what sticks. As the industry has started to mature, standards evolving, emergence of new channels, formats and walled gardens, the utilization of learning systems AI is prime for the next wave of development; and within this realm exists experienced and nuanced ad traffickers and media planners which can clumsily navigate multiple DSPs, Facebook campaigns, search campaigns, and interpret objectives, results and success. The increased complexity of the systems involved, scale of data, and response times required to be efficient and successful, the human powered ad technology is endangered, though I would argue that the human assisted AI for ad technology will still be viable and valuable for quite some time.

Still the biggest problem in most industries and applications is "what to do next". This is where machine learning and AI holds it's greatest promise.

Uber: The Good and the Bad

noreply@blogger.com (Paul Neto) — Thu, 21 Apr 2016 21:10:00 +0000

I am fortunate that with my occupation I get to visit many cities and interesting places across the US, Europe and Mexico. I've become a regular user of Uber and spend way too many hours in lineups whether at an airport, taxi stand, or hotel. As such, it is of paramount importance for me to be efficient in my modes of transit.

I can consider myself as an early adopter of Uber from the beginning, and am often annoyed when I land in s city where Uber is not available. Each time I enter an Uber I make a point of speaking to the driver to get their feedback on the service, their level of satisfaction, and how they perceive their role in the transportation industry.

In large, Uber has been perceived by drivers and the general population, or at least my peers and immediate circle, as a positive experience and service. For many drivers it has helped them achieve a work-life balance, or to fill the gap while they are pursing other dreams (think entrepreneurs, actors, musicians), or as a new career.

While Uber and competitors have been growing in popularity, in many places it has not been accepted with open arms, including legal battles, and even violence. The most interesting aspect of the Uber revolution is how it is helping redefine business models and challenging an industry that has not innovated in decades. It is a model that while may be disruptive in the short and mid-term, will experience major iterations of fine-tuning over the next five or ten years, and evolve into something sustainable and common-place.

Though I do not believe that Uber will work everywhere, or should it. Let me provide a few examples.

The promise, and premise of Uber, in the early days at least, was to optimize use of all the black limos that sit idle for hours per day, and provide customers with a frictionless way of getting around. The most painful part of taking a taxi was hailing a cab, waiting in the rain hoping one would drive by, then fumbling with cash and coins and trying to figure out an appropriate tip. All in the meantime, hoping not to get ripped off and taken on a scenic ride. It is important to note here that the notion of cost wasn't even part of the equation. It is about the transaction and delivery of service.

The free marketplace is intended to permit supply and demand, for the adoption of new products and services as the opposing forces to regulate success and failure.

The first classic use case for justifying the existence of Uber is in cities like New York, Toronto and Lisbon. In these cities, from my experience at least, is the classic example where it can be difficult to hail a cab, where many cars are extremely dirty and smell, where drivers will ask you where you are going then say no, where many drivers are rude, and there is little confidence that they won't take you for a longer than necessary ride. I will be clear that this is not every driver and not every experience, but it is true in enough incidences that it spoils it for all the good, honest, hard working drivers out there. The working conditions for many of these drivers simply would not be tolerated elsewhere. In some of these locales, the cars and 'badges' are owned by monopolies, and is a very hard industry to make a living.

In these cases, the disruption by ride-sharing services is completely justified. There is a product/service that is superior to the existing, it is a better deal for the driver/employer/contractor, and has value for the customer. This should be a happy win-win situation for everyone. In this case, the system has worked.

Use case number two. Traveling to unfamiliar places, some of which may be notorious for having unreliable or unsafe taxi services, and many locals stressing to use Uber instead of local taxis. These technology driven services in these cases adds a level of accountability. One can travel with minimal cash, the transaction is paperless, and completely tracked using the technology infrastructure. For the weary traveler, this peace of mind and convenience is well received.

Though, as I started out, this system isn't meant to work everywhere. A prime example is London, England. The iconic black cab in London is nothing like that in say a New York or Toronto. While the system is highly regulated, the premise is that anyone who drives a cab and can pick up someone on the street to drive them somewhere, is essentially an ambassador to the city. They carry a lot of responsibility and as such, should be held to very high standards. As a result, as a driver in London, you must take a 3 year study called the "knowledge' which is a grueling study on every inch of London. Need directions, need advice, need a good recommendation of a place to eat, find something out of the way, or the best local spot, there is not a more helpful and knowledgeable resource than the nearest black cab. The black cabs may be pricey, but there is value to what they do and plays a very important role in the operation and culture of the city.

Is there value to protect this service, to protect the culture of getting around London. I believe there is. The challenge is in the balance as Uber has threatened this way of life.

These are just a few examples of they benefits and complexity of a changing business model. Though while many of the proponents of the disruptive business model point to the free market capitalist forces as part of the justification, there is a social and moral obligation as well. As much as we all like to think we operate within a free market model, there is a very strong social reality and responsibility. We do not operate in a void of letting the markets take care of everything. We are a charitable society, one that values prosperity, health and general wellness.

My point here is that there are very few occupations where someone with few qualifications, new to the country, or less educated, or in a current bind where they need to make ends meet, can take employment and make it through the week. Some will say Uber does this if fact, and better than the taxi industry, though it does not eliminate the fact that the current taxi industry plays an important role for many today.

The bright side is that the ride-sharing, or on-demand economy is still in it's early days and will continue to evolve and iterate. In a few years from now, it'll be very different.

5 Ways Californication is Like Seinfeld

noreply@blogger.com (Paul Neto) — Tue, 13 Jan 2015 15:19:00 +0000

Like most of us, I've been a long time fan of Seinfeld. There is no doubt that Seinfeld will go down as one of the best sitcoms of all times. Thanks to Netflix and other streaming servies, like everyone else, I've done my share of binge watching. Of most recent is about six seasons of Californication with David Duchovny.

At first glance, both of these shows are quite different though there is something that many of us find appealing about both. Of course Californication is much edgier, largely due to it's home on Showtime, and Seinfeld being a primetime show during it's run. I've often wondered what Seinfeld would have been like if it was hosted outside of a major network like NBC? Some say Larry David's Curb Your Enthusiasm is what Seinfeld would have been outside of the primetime slot.
While both these shows contrast in some very dramatic ways, I believe what makes both of them so likeable are the ways they are similar. Here are 5 ways Seinfeld and Californication/David Duchovny are strikingly similar.

1. Witty Humor
Both Seinfeld and David Duchovny use witty, quick-fire humor primarily about obvious observations around them in any situation

2. Not so Normal
The characters surrounding Seinfeld and Duchovny are incrementally more dysfunctional that we think they are the normal ones. Each on their own without the context of the supporting characters would be a different story.

3. Guest Stars
Both shows strategically use well placed guest stars. Whether it is Rick Springfield in Californication or Keith Hernandez in Seinfeld, all are well placed and add a special dynamic to both shows.

4. Bi-Coastal Madness
Where Seinfeld is about the neurotic New York, Californication is about the over-the-top LA. Both draw on stereotypes of each place and use established landmarks. Interestingly both venture to the other coast where the characters exhibit new levels of dysfunction which drives them back to their home coastal city.

5. It's Really About Nothing
Not surprisingly, both shows are really about nothing.

The Under-appreciated corner number 4 CTMP Mosport

noreply@blogger.com (Paul Neto) — Sun, 13 Jul 2014 02:41:00 +0000

Corner number four at Canadian Tire Motorsports Park, formerly known as Mosport, is the often overlooked and under-appreciated corner. Corner number two, it's double apex, blind corner, odd camber, high speed sweeper gets all the drama and attention. On the contrary, corner number four is probably as exciting if not more important for an aspiring driver.

Most drivers will focus on number two starting cautious and building up momentum and confidence, and lap after lap prepare for it again. Corner number four is as exciting and important for a number of reasons. It is also a great opportunity, particularly for those of us in relatively underpowered cars in

the field. Having an elevation change of about 90 feet (corner number two is about half that), and despite how it looks has camber in your favor which means tons of traction. You can only appreciate this if you have an opportunity to walk the track. What this means is that once positioned correctly you can easily power through the corner gaining tremendous amount of speed in a very short amount of time. It is a great opportunity to make up some distance on those Carrera S's and Turbos.

Secondly, this importance is amplified considering what is coming up next. Turn five A and B. Any way you slice it, turn five is the slowest part of the track. You either try and crank through it or just accept and sacrifice it. As such any progress you can make coming into it (turn four) will give you an advantage. Another benefit is that the approach to corner number five has an incredible steep rising slope. Turn five is so steep that when walking you must climb on your toes. This plays in your favor as you power through corner four and delay braking to the point that it feels too long. The incline of corner five will help you compress and amplify braking power.

As you come out of corner three on the far left side, ensure you approach corner four far to the right. Approaching the crest be sure to turn in early and align so that you are turned in and passing under the Continental sign approximately in the middle. Look far ahead and your goal is to be positioned and aligned tight to the left of track at the bottom of the hill. It may not look like you will get there but hold steady and as soon as you are over the crest, power down and the car will rotate as you come through and align perfectly so you are tight to the left edge at the bottom of the hill. With the elevation change and positive camber you will experience great acceleration and continue accelerating until it feels almost too late. Heel-toe technique is quite important here as you need to shift down quickly as you brake hard. The change in slope will assist in braking and engage as coming in to turn five.

As you build experience through this corner, you will be amazed as you can consistently build time, speed and power. Master this corner and number five becomes is merely a setup for the Mario Andretti Straightaway.

Turn 2 (CTMP Mosport)

noreply@blogger.com (Paul Neto) — Mon, 23 Jun 2014 03:19:00 +0000

A couple years ago I took up performance driving as a hobby, so I picked up a little Porsche and joined the Porsche Club of America, Upper Canada Region who put on a great program with their driver education events. I really have no interest in racing but am enticed with learning about the dynamics of speed, grip, and techniques of maneuvering a vehicle around a road track.

Living in southern Ontario we have a little gem in our backyards called Canadian Tire Motorsport Park,

or as once known Mosport. Once a regular stop on the F1 circuit, it was neglected for many years though in the past few years Canadian Tire has transformed it into a modern facility. Often regarded as one of the most difficult courses in the world, it is one of the few circuits that has largely been untouched and mostly original. Existing in the oak ridge moraine, it has earned it's reputation due to it's high spend turns, blind apexes and dramatic elevation changes.

Turn 2 is probably it's signature turn and once you drive it, it's like playing a spectacular golf course where you replay certain holes over and over in your mind. Turn two is a double apex left turn sweeper and driving it is best described as a dance with physics, grip and technology. Check out this youtube video of turn 2.

Coming out of pit lane you want to quickly come up to speed as taking turn 2 too slow makes it an awkward turn to maneuver. As you build experience you start to find the groove and speed where things just start to work and you know this when you feel it, no other way to describe it. A quick lap or two around the track to warm up the tires and turn two comes to life.

Coming out of turn one you want to be flat out when you hit the apex and gradually line up to the right to approach turn two, making sure to keep clear of the pit merge lane. Approaching the turn there is a crest which makes it a completely blind corner. The trick to this turn is to stay committed and keep your hands still and keep to the line. As you approach, lightly tap the brakes to shift weight to the front and load traction on your front wheels and while still in the blind you want to make your turn in. The best way to describe the amount of turn in is that if you were going any slower you'd probably drive off the track. At this point you are committed with your turn in and you will quickly see the first apex as you come over the crest. My preference is to actually put my left tire in the gutter, yes, in the gutter. Your speed if done right will float you nicely and put you in position.

Coming through the first apex you are feeling that the road slopes away from you. Since you are sweeping to the left and the car and having just come over a crest there is a tendency for the car to unweight. This is not good and 99% of all spins in this turn come from those not managing this correctly. To avoid being unweighted and to maintain traction, you must keep some level throttle through the turn. The worst thing you can do at this point is come off the throttle which will cause the car to lighten and the back end will start to wiggle out. Just check out YouTube videos of "mosport spins" and you'll find a few of these.

Holding tight, and if riding on street tires, you will feel tire technology at work as the tire rubber is peeling at the road. The car should naturally track out to about mid track and apex two is in sight. You know you've timed and lined up things nicely as at the bottom of the turn you will feel the car actually sit and hunker down as you tighten up and literally launch towards apex two. At this point traction is plentiful and you apply more throttle to come out of turn two as fast if not faster than when you came in. At apex two you should be full out on throttle and let the car track way out to the right using up all the road.

If you need a little guidance in turn two, luckily two years ago there was a patch of pavement at the bottom of the turn approaching apex two that was re-paved. If you line up with this, you've done well.

A blind corner can be intimidating and even if you miss on the turn-in for apex one, don't worry and don't fight it to try and get back in line, just focus and line up to hit apex two correctly and you'll have plenty of speed coming out. Remember, slow in, fast out.

How Social Media Saved Online Market Research

noreply@blogger.com (Paul Neto) — Wed, 12 Mar 2014 18:15:00 +0000

There was a period of time roughly between 2004 and 2010 that I describe as the dark ages for online market research, or more specifically, for anything related to online surveys in general. This was a period of time where the rise of behavioral analytics and a good few years of rapid growth and questionable practices within the industry that put a dark shadow on online surveys.

The market research industry really needed to be held accountable for the mess. Online surveys had become rampant, online panels were growing at an accelerated rate, new panels emerging over night, ridiculous panel acquisitions were taking place, some going public, and the overall practice was questionable on so many fronts. Panels were built by offering people survey points, free iPods, and by downloading spyware through screensaver downloads. The amount of overlap between panels was huge (40%-60% in many cases) and the number of surveys per respondent was increasing.

This ecosystem gave rise to what was known as the professional survey taker. During this period, it seemed that every market research conference was focusing on panel integrity and understanding the professional survey taker. Every talk had the same result. There was inconsistency in results across panels, and within panels.

The core problem with these so-called professional survey takers is that 90%+ of surveys were being taken by less than 1% of the audience. Secondly, due to their behaviors, it was difficult to account for these survey takers to be representative of anyone. The typical survey taker was low income, middle America females, they spent unhealthy amounts of time online, collected coupons, and clicked on everything. Effectively, it was almost impossible to find a nationally representative cross section of people online for research purposes. It was clearly understood and excepted that the online world was not representative of the real world.

Then something changed . . . and what changed was social media arrived.

Once social media arrived, which included Facebook, Twitter and others, the notion of sharing, clicking, providing feedback, comments and other types of user directed contribution, fundamentally changed the profile of the online individual. Facebook quickly grew out of being an early adopter nice group of college students, and quickly there were as many 55+ as there were high school students on the social network. Oprah helped propel twitter into the mainstream and soon it became the exception for someone not to be engaged online.

More than ever people become willing to provide feedback, share their thoughts, post pictures, and more importantly felt they had a voice and that their voice counted. While social media quickly became noisy, online market research surged with the ability to generate consumer feedback in a structured and quantified manner.

It has become very common for media channels to publish survey research results as they quickly found that the data was easy to collect, and readership finds this format easy to digest and very interesting. It is very simple and effective to tweet that 17% of people have sent a tweet while in the bathroom.

Where survey based research was once considered antiquated, not representative and boring, more and more companies began adopting it, or simply showcasing it, as a viable and valuable data source. Of most interest was Google's entry into market research. Yes, the company that stood proud on it's algorithm heavy approach entered the survey market in 2012 with it's Google Consumer Surveys product. In 2012 statistician Nate Silver ranked Google Consumer Surveys second in accuracy over many long time standards such as Gallup, Ipsos, Angus Reid, and CNN during the 2012 Presidential election.

Net result is that the online market research is alive and well and has established itself as a viable data methodology. Thank you social media!

The annoying Roku Error Code 014 and 013

noreply@blogger.com (Paul Neto) — Tue, 16 Apr 2013 16:33:00 +0000

My Roku 3 arrived yesterday and was eager to hook it up. Unfortunately I experienced the ever annoying error codes 013 and 014 when trying to connect it to the Internet. Browsing all the forums and the best that Google could do did not solve my problem. As a result I ended up calling their support line and they gave me the instructions to access the 'secret' screens.

The problem is related to outdated software on the Roku which restricts it from connecting whether via wireless or wired.

1. Update Software: On your remote please press the house 5 times, press the fast foward 3 times, then press rewind 2 times. You will get a secret screen, There you have to check your IP address that it is not all 0's. Select the option to update software. The Roku player should now restart.

2. Disable Ping: Once that the roku finish the set up, on the remote please press the Home button 5 times, then the Fast Forward (>>) button 1 time, then the Play button 1 time, then the Rewind (<<) button 1 time, then the Play button 1 time, and then hit the Fast Forward (>>) button 1 time.

This will take you to the Platform secret screen. There you have to find the option to disable network ping. Select it and go back to the set up process of the roku player.

That should do it.

Annoying but works.

Where Market Research and Big Data Intersect

noreply@blogger.com (Paul Neto) — Mon, 10 Dec 2012 22:25:00 +0000

Market research, or at least survey based market research, is often referred to as Little data. Big data is the hot topic these days and is slowly infiltrating all industries. The discussion today is around the meeting of Little data and Big data and how the two complement each other. The definition of big data varies greatly depending who you ask.

Big data as I describe it, is a data source that meets some of these criteria:

The quantity of data is great. Managing, accessing, and processing is a challenge
The format is often unstructured, or at least seems to be unstructured
Data source is very complex
The user just isn't very sure what to make of it

Little data, survey data, or market research in general is most efficient at adding a human interpretation to data. Some call this the 'why' or the 'insights'. You may wonder where do these two data concepts intersect.

A powerful component of Big data is machine learning. This field of expertise is essential in helping blend the two concepts of Big and Little data. For any of those interested in this field, I strongly recommend spending a few evenings or weeks listening to Yaser S. Abu-Mostafa's lectures on ML available through iTunesU.

As explained by Yaser, machine learning is a candidate as a technique for any problem where:

Some pattern exists, or we believe one to exist
We cannot easily describe this pattern mathematically
Data exists

Some of the challenges in MR are related to the changing data sources. Whether it be text analytics, social media, web metrics, or other, the vast amount of data being made available is overwhelming. Excel, Word and general databases are no longer sufficient to dealing with the data we must manage.

Machine learning is a technique that I believe MR firms will need to adopt in-house, or through strategic partners to be able to deal with the data challenges they face today. This would surely help assist reducing data to the most essential metrics that market research can apply it's magic to.

My prediction is that machine learning will play an important role to define the intersection between little data and big data.

The 7 Essential Web Services for EVERY Startup

noreply@blogger.com (Paul Neto) — Fri, 23 Nov 2012 17:51:00 +0000

Start-ups are fun, exciting, scary and should be on everyone's bucklist (at least once). It's helps build endurance, a thick skin, some real street smarts and a real appreciation for all the little success (and horror) stories around us. A key concept is to keep things moving forward, to build momentum. Each step along the way, how miniscule it may seem, whether an email, a phone call, a note, or a line of code, all in aggregation help contribute to the success.

Thus from day one simplicity and productivity is critical. I've put together a list of web services, apps or whatever you want to call them, that I believe or at least have experienced to be essential for any and every startup to help maintain focus, communication, simplicity and productivity. In no particular order here they are.

Basecamp. From the clever team at 37signals, Basecamp is a brilliant project management application. Effective for managing and tracking projects with disparate teams, or collaboratively with clients this is a must.

Google Apps. This is an obvious one but the easiest and most inexpensive way to get things real with email, basic document management, collaboration, wiki etc.

Pivotal Tracker. Great tool for agile software development teams for tracking and managing projects, requests, bugs. Also extends well into managing support requests.

Hootsuite. Every startup needs to manage their social presence. Hootsuite is a good standard to launch with.

Campaign Monitor. Great design, intuitive and feature rich, Campaign Monitor is by far my favorite service for managing subscriptions, email lists, newsletters and email outreach.

Campfire. Another beauty from the 37Signals team and a must for any distributed team. Think of campfire as a group chat with persistence. It helps build team dynamics across globally scattered teams, and general chatter to keep people connected. We've found it useful even when on vacation as a great place to check in a scroll through the daily logs.

Dropbox. File sharing can be a real nightmare. Dropbox makes it easy and a life saver when on the road. New team options makes it even better.

There you go. All are very cost effective, and will make your teams much more productive and engaged.

Where will the advertising dollars go after DNT?

noreply@blogger.com (Paul Neto) — Mon, 12 Nov 2012 16:57:00 +0000

The sides are divided on the Do Not Track (DNT) issue since Microsoft decided to enable the DNT header in Internet Explorer 10. After feedback and action by the W3C, Apache, the DAA, and now Yahoo claiming that they will not honour the DNT track signal in the IE 10 browser, the entire DNT initiative is in jeopardy. This is a real shame because at the essence of the initiative is consumer privacy.

While some (myself being one of them) don't believe that the DNT setting should be enabled by default, I do believe that all should abide by the setting to illustrate that privacy in the industry is taken seriously. The last thing we all need is the consumer (they key individual in this entire debate) to become more skeptical about privacy, privacy policy, and privacy compliance. The uninformed consumer may only understand that privacy settings are being ignored. That is the wrong message.

It is also time for those who are predicting the death of online advertising if DNT persists, to acknowledge the resilience of the marketplace. Latest numbers indicate that online behavioral advertising accounts for only about 15% of the total ad spend. While 15% may be considerable, it does not signify the collapse of an industry if it was to disappear. A few points to consider here:

The negative impact, if any, at most will be a fraction of this amount
Ad budgets will not decrease, as such, dollars will simply be reallocated to other forms
These pressures will spur creativity, new technology and services
There is an argument that a scarcity in behaviourally targeted inventory may contribute to increased quality and pricing of the available targeted inventory
TV, print and radio advertising have been around for years and are not dependent on tracking

Do Not Track and evolution in consumer privacy policy and expectations will surely have an impact on the ad supported web and associated industries, but it is also true that the players in the field will need to evolve. If the end result is that behaviourally targeted advertising ceases to exist due to technological changes and consumer demands, then be it. I don't see many people upset about the demise of the VHS rental industry or others that have evolved out of existence. In the meantime, it's business as usual.

@paulneto

5 reasons I upgraded to the iPhone 5

noreply@blogger.com (Paul Neto) — Fri, 12 Oct 2012 18:24:00 +0000

Yesterday I picked up my iPhone 5 and quietly powered down my iPhone 4. I chose specifically to avoid the 4S as it was a marginal upgrade in my opinion. Though I am quite happy with my new device. Here are the five reasons why.

This thing is feather light. When it is time to retire this device, it will barely be worthwhile as a paperweight
Slim factor. A bulky smartphone in your pocket is irritating and I'm a minimalist at best.
Taller screen. The screen is a dramatic improvement over the iPhone 4 not only in resolution (the retina display is gorgeous) but also in size. Now nearing 16:9 ratio, the height also permits more app icons per screen and per folder. For those heavy folder users (for those numerous of apps added by your kids) this is useful
Speed. I recommend you line up each generation of phones and fire it up, launch and app, and you'll see how fast it really is.
New maps. Yes, I'm really including maps on the list. Living in Toronto and frequently travelling to NYC and the bay area, data usage and map access has been a nightmare over the years. The one thing the new maps does extremely well in caching of the data. I can load a map in Toronto, land in SFO and zoom, scroll, pan with little need to connect for data refresh. This to me is huge and of all the ones on this list, the most striking for me.

Let's imagine web tracking went away tomorrow

noreply@blogger.com (Paul Neto) — Mon, 23 Jul 2012 19:40:00 +0000

In the last few blog postings I've been discussing various aspects of the Do Not Track (DNT) issue and third party tracking online in general. I think it would be interesting to take a look at this issue from the perspective where third party tracking was to go away immediately. One can only postulate what the impact may be but I think it is worth a discussion.

To set the scene let's assume that the pressures of the EU directive on privacy, the do not track debate, and third party tracking for purposes of delivering relevant advertising has led to 50% of online browsers having DNT enabled. Here are a few potential outcomes:

Small publishers perish Thousands of small publishers immediately cease to have a viable business model. This is unfortunate but I don't hear anyone crying about the loss of thousands of video rental stores either. Sometimes business models change, some survive and some don't.

Google. Stock tanks.

Lumascape landslide. Many of the data and service companies on the display advertising Lumascape crumble, are absorbed by others, or simply cease to have a viable business.

Ad technology investment. In conjunction to unraveling of the existing display advertising lumascape, there is an increased level of investment and opportunity in the ad technology space.

Premium publishers rejoice. For years the premium publisher has been trying to fight off pressures of declining CPMs, loss revenue, data leakage and third parties marginalizing their audience. The time and investment into curating content, building audience and audience based insights starts to pay off. These publishers know the most on a first party basis who their audience is, are able to mine the data, provide real insights and deliver services to both advertisers and viewers.

Tracking becomes more valuable. As the availability of individuals tracked online declines dramatically, the shortage in supply makes them much more valuable to advertisers looking to get relevant ads in front of individuals. CPMs of targeted ads increases as a result.

In all cases, a primary point is that online advertising ad dollars probably will not diminish or disappear. While there may be a bit of a stall while things re-adjust, dollars will continue to flow, though the flow path may shift from time to time. Isn't this what should happen in the markets?

Imagine if research was free?

noreply@blogger.com (Paul Neto) — Wed, 27 Jun 2012 14:12:00 +0000

"Imagine if market research was free and clients only paid based on performance?"

--- Paul Neto, just making shit up

DNT is a wake-up call. Stop and listen

noreply@blogger.com (Paul Neto) — Tue, 26 Jun 2012 14:52:00 +0000

Much of the debate around the Do Not Track (DNT) controversy is around Microsoft setting it as a default in Internet Explorer 10. The debate has been a healthy one for the industry and Microsoft deserves credit for taking a stance. Often it is the case that change only happens when an individual takes a position and is challenged. I don’t believe anyone is truly against the DNT specification, but not everyone is on the same page on the ethics of using it, the default settings, and the consequences.

Though I will point out that everyone should take a careful note and watch how this unfolds. Do Not Track is a wake up call. It is a wake up call for consumers and for marketers, agencies, technology companies and even the investment community (VC's, angels, founders etc.).

A few things are a definite:

privacy is taking front stage with consumers
consumers have and will continue to have more options around privacy
privacy is getting the attention of policy makers
the EU directive has broken new ground and everyone is watching closely
tracking and privacy is not going away and will only become more scrutinized
ad targeting will be very different than what we know today
we have an industry that is anchored in dated technology

Yes it is possible that the behavioral targeting/advertising is at risk of going away at the flip of a switch as stated by Tom Espos who calls for a BT contingency plan. In reality, this of course will not happen over night, this of course won’t cause a lumascape landslide over night, and web sites and the internet will not crumble. It will be different, but it will not be as catastrophic as some may portray it. Some will suffer, some will die, some will change, and many will emerge and flourish. Whatever the outcome is, this will be for sure.

Your concern should be where you fit in all of this. Are you ready for change? Let’s for a moment consider what behavioral ad targeting is riding on. It’s not the threat of DNT or other privacy initiatives, it’s biggest threat is it’s reliance on the third party cookie. At the crux of the issue is this little third party cookie that was probably intended to be used for convenience by web sites and services, has been abused, poorly implemented, and circumvented a whole industry.

If I was involved in any cookie based dependent technology, I would be scared. Scared at the fragility of it. Here are a few points to consider:

third party cookies (or cookies in general) are very unreliable
they get deleted all the time
browsers vary greatly on how they manage them
user settings are all over the board
have a bad rap in the media
have a bad rap by most consumers
are very limited in the data they can store
the size and number of cookies are inconsistent across browsers
are awful for any measurement
multiple users across machines renders them mostly useless and reliable

For these simple points, ad targeting cannot not evolve as a third-party cookie based technology. Those who survive and thrive will be those who can see beyond today’s limitations and evolve their technology and services.

For us at Crowd Science, (warning: shameless plug for the company I am a co-founder at) we specifically and purposely built our technology stack around opt-in using research principals, around value propositions for the publisher and visitor, and without the reliance on cookies. Crowd Science Audience modeling is based on a sophisticated audience prediction engine, and only utilizes cookies for convenience, not as a necessity, and our little know secret is that a cookie-less operating mode has been in operation since the inception.

Is DNT (Do Not Track) an opportunity?

noreply@blogger.com (Paul Neto) — Fri, 22 Jun 2012 19:08:00 +0000

There has been a lot of discussion and controversy in the media recently over the announcement by Microsoft to implement the Do Not Track setting (DNT) as a default in the next release of Internet Explorer. It's difficult to gauge the actual impact this would have but Tom Espos puts it well in his blog posting What's your BT contingency plan?

"It’s possible for behavioral targeting to go away with the flip of a switch.
That switch can take the form of legislation in the U.S., a popular browser plugin, new browser technology or in any number of other forms. What form it takes doesn’t matter as much as having a contingency plan in place, just in case some of your favorite targeting methods can no longer be implemented, thanks to new technology or new laws."

While this issue will need time to pan itself out, those who are thought leaders and innovators will attempt to see beyond this current crisis. Of note is John Battelle who poses the possibility that DNT is an opportunity and not a threat.

"To me, this presents a huge opportunity for the owner of a site to engage with its readers, and explain quite clearly the fact that good content on the Internet is paid for by good marketing on the Internet. And good marketing often needs to use “tracking” data so as to present quality advertising in context. "

For those who are involved in some type of behavioral targeting (BT), whether as employment, trying to build a business, or even an investor, you should be worried. Coupled with the precedence set by the EU Data Protection Directive, you can be sure that this is not going away.

The most important metric in audience based ad targeting

noreply@blogger.com (Paul Neto) — Fri, 13 Apr 2012 01:31:00 +0000

Online advertising is evolving at an alarming rate. Currently there is a shift towards audience based ad targeting as outlined by the eMarketer report. Though when it comes to ad targeting, there is a metric that is often left out of the conversation. This metric is the in-segment incidence. The best way to illustrate this is with the following graphic.

Consider each dot to be a delivered impression. Based on the targeting, or lack of targeting, there will be an incidence of a given audience that you will hit successfully (this your target audience). Consider each orange dot an impression delivered to the correct audience. In few cases are you ever successful 100% of the time.

With the plethora of data sources available, whether third party or first party, there is little transparency to what the true success rate is of reaching a given audience. Further, there is even less transparency of what the expected success rate is. For instance, there have been some yet undisclosed tests where a segment from the third party data provider Blue Kai experienced an in-segment success rate of 11%. That means that for every 100 impressions, only 11 are reaching the intended audience.

There are a couple important concepts to take note of:

Natural Incidence
Every audience exists in some concentration naturally. Whether it is an audience of wealthy females, or tweeting dads, most publishers will have some natural incidence of finding these individuals on their sites. This may be considered as the base or an index of 100. There is also a distinction between the natural incidence on a particular site, and the Internet Norm. For the purposes of this discussion, we are interested in the incidence on the particular web property.

Targeting Incidence
The targeting incidence is just that, the incidence that targeting that audience targeting would provide. In an ideal world, audience targeting would provide an in-segment incidence of 100%, though this is rarely the case, and in more cases, less is acceptable.

Targeting Lift
A measure of performance in ad targeting is the lift in reaching a desired audience. More specifically, the percent change of the targeting incidence over the natural audience incidence. Lift plays an important role as it provides a relative improvement.

Consider the situation where a certain audience exists on the Internet at an incidence of 5%. If a web property, due to a variety of reasons, in fact has a natural incidence of 20%, it is effectively 4x more effective in reaching the defined audience. At first glance this is a very viable and valuable segment with absolutely no targeting (ROS). Introduce the CITRUS ad targeting technology, the targeted incidence may be as high as 80% in-segment. (yes this is my shameless plug)

As noted, there is little transparency in the industry around what in-segment means and what acceptable rates are. As part of the Crowd Science CITRUS model, a core component is understanding the three concepts above and providing the publisher with the insights on expected performance, which in turn, translates in to value.

In my next post, I'll dig in to a few insights around some industry related performance metrics.

The notion of 'must have' in sales is no longer valid

noreply@blogger.com (Paul Neto) — Wed, 11 Apr 2012 17:50:00 +0000

One common tactic in the sales process is to emphasis or create a notion of a product or service being a must-have. There are a number of tactics used in sales to create this tension and help prospects over the line. I won't go into these here as it is not very exciting nor interesting.

Often the question is to look at the product or service you offer and justify it's value based on it being a must-have. The reality is that this idea of a product or service being a must-have is old, outdated, and simply wrong. Other than some very basics, and to be clear bottled water is not a must-have, there are limited must have products and services in this society. Plenty of want-to-have but for every product and service there is an equivalent and suitable alternative.

Price is not the purchase driver it used to be. Take a look at Apple products, in many ways inferior and overpriced, yet their sales volumes are ever impressive. Trying to go to market with a new product or service? If you plan to use price as a motivator, unless you are prepared to lose money, there already is or will soon be someone who can provide it for 5 cents cheaper.

Successful sales comes from building a relationship/connection with your clients and prospects. It comes from people understanding why you do what you do and what your vision is. Everything else is secondary. Sell and communicate based on the why and you build a partnership and an alliance. Establish either of these and the end result is a partnership where two parties want to work together. At that point it's simply about aligning of opportunities and timing.

Why do people buy Apple products, because Apple builds cool products and consumers want to be part of the culture. The purchase decision is made based on the 'why' and not the what. Is it a must-have, absolutely not. Are there alternatives - absolutely.

Google enters survey sampling (finally)

noreply@blogger.com (Paul Neto) — Wed, 04 Apr 2012 22:15:00 +0000

Google recently announced it's entry into the survey based research world with it's Google Consumer Surveys product. This has been long awaited and while much of the reaction has been around it's threat to the DIY survey tools market such as SurveyMonkey, the real impact is on the online sampling market.

Market research has long been a poorly engineered industry. It is repetitive in attempting to do what it has always done poorly. For example, surveys by the industry have really not progressed in the past 60 years. With the introduction of the Internet, instead of innovating, the industry simply took it's old paper survey and marked it up with HTML and called it innovation.

Sampling hasn't improved much either. Consumer panels have existed from the beginning and with the internet they simply just became bigger. To improve panels they aggregated panels and literally make the process destructive to the credibility of the industry.

The real importance of the Google entry is that sampling and panels are being challenged. Secondly, market research for once is being defined by it's value proposition. This value proposition being value to the publisher and the respondent. Forget the use of survey points and win an iPod contests.

And for all those market researchers out there that will look down at Google Consumer Surveys due to not supporting 60 question surveys, I must say that it is in fact those 60 question surveys you are running that are most harming this industry.

What I find most interesting in all this is that Google is now acknowledging that attitudinal and user feedback data is of value. A sign that the Internet is maturing beyond strictly behavior and watching.

Is Pinterest the Anti-Facebook?

noreply@blogger.com (Paul Neto) — Tue, 28 Feb 2012 19:41:00 +0000

Pinterest is quickly becoming the fastest growing social app/network in terms of adoption. It is rather interesting as well that the growth of Pinterest is largely due to women (some say 80%). My theory is that Pinterest is establishing itself as the anti-Facebook. Here is my take on this:

Pinterest is very visual. Facebook has become too busy, too much content and text
Pinterest is naturally about exploring. Facebook is about updating
Connections in Pinterest are very light which is definitely the trend and has a very different mindset where Facebook relationships are much more restricted
Pinterest is about creativity and can capture the attention of an individual like a full page ad. You know that moment where you hesitate and take it all in. That's it.
Pinterest is akin to the grocery store checkout aisle trashy magazines. These are those full color, flashy magazines around celebrities, fashion, weddings and everything else
Pinterest is about branding and discovery. Branding is where the dollars are. Facebook nor Google will never be successful at this
Friends get annoying. Pinterest is a good distraction from Facebook.

Every 60 seconds . . .

noreply@blogger.com (Paul Neto) — Tue, 14 Feb 2012 20:09:00 +0000

Waze will change how you get around in Traffic

noreply@blogger.com (Paul Neto) — Tue, 14 Feb 2012 12:23:00 +0000

Waze is a start-up app to watch out for. Transforming how you get around in traffic.

The Lack of Transparency in Behavioral Ad Targeting

noreply@blogger.com (Paul Neto) — Mon, 14 Nov 2011 15:42:00 +0000

Behavioral and contextual audience ad targeting is experiencing a healthy boom in the industry. Though while there is much growth, perceived and real success, the lack of transparency in behavioral ad targeting is a real risk to the validity of the methodology. Today many data companies are offering high value in-market segments at improved CPM rates; but do consumers of this data really know what they are getting? Do they really know the quality of the data they are purchasing?

The best way to illustrate this is to take an example. With behavioral targeting, site visitation is a primary proxy for an individuals interest. For example, the following online behaviors may classify someone as an in-market car buyer or car enthusiast.

Visitation to a car review site
Configured a car with an auto online configuration tool

While this may be very true and valid, let's look at the example a bit further. More specifically, let's consider these as:

visitation to http://www.audi.com
User configured an Audi R8 on the Audi site

If using the criteria further above, both of these scenarios classify a visitor as an in-market car buyer or car enthusiast. At first glance this may be a safe proxy for a high performance in-market car buyer. After all, configuring your car is a great indicator that you are ready to purchase an Audi R8. Wow, this is some valuable data. But let's not get ahead of ourselves. As it turns out, the Audi R8 is one of the most popular car configurators on the web. This leads me to wonder whether these are really in-market car buyers or actually in-mid-dream-car-dreamers. Let's face it, a car configurator, and in particular one for the Audi R8 is real car-enthusiast porn.

The point here is, if you are a data buyer for ad targeting purposes, demand your supplier to disclose on what criteria do they use for classifying their data, and further, at what incidence level should you expect these to occur with your life targeting. If you can't get a clear and quantifiable answer, I would suggest this raises some real question on what you are really buying.

Adding Social Targeting to Audience Targeting

noreply@blogger.com (Paul Neto) — Fri, 11 Nov 2011 20:54:00 +0000

As reposted from the company blog at http://blog.crowdscience.com/2011/11/adding-social-targeting-to-audience-targeting/

Emerging research is building a case for the integration of social media into ad targeting strategy. This evolving area of targeting is referred to in a few ways, namely social media targeting or just simply social targeting. Social targeting is involves targeting ads to users based on their social behavior, such as Facebook ‘likes’, posts and Twitter tweets. Social media data is an ever evolving array of fragmented data. I’ve commented earlier how social is a very noisy medium and everyone is struggling to decipher some signal out of the noise.

Long are the days of generic ad targeting, or even further generic audience targeting based on a few core demographic characteristics. But how can we integrate social targeting to the audience targeting world? One approach that we are taking here at Crowd Science is to use our research methodologies to build an online social persona. These personas may take many shapes, including categorizing groups of individuals as social influencers and social addicts. The premise here is to evolve these various personas while attempting to gain insights into their habits, behaviors, motivators and drivers. We can then utilize this data to turn these profiles into targetable audience segments.

We are driving this a step further by providing the ability to build very specific micro-segments such as ‘connected students’, ‘social influencer technophiles’, or ‘online social addicted grandparents’. Using this kind of hypersegmentation provides audience segments that have exponentially more value to advertisers than traditional methods.

Below is a sneak peek at what we’re going to be adding to our SEGMENT + TARGET components of our CITRUS platform in the next feature release. This new functionality allows online publishers to take any of their existing segments and socialize them using this model. It is a just sneak peek (we had to hold a little back to keep our marketers happy), so drop us a line or keep us in your twitter feed to get updates on the socialization of audience targeting.

Responsibilities of a Manager in a Startup

noreply@blogger.com (Paul Neto) — Wed, 11 May 2011 21:52:00 +0000

The people you work with is one of the main things that makes being involved in a startup so exciting. Everyone becomes like a family and deeply involved in a vision. With so many ups and downs, there are many battles won and much pain endured.

People lean on each other for support, encouragement and advice through each days battles. Anyone who has ever been involved in a startup environment can attest to this. Particularly for anyone who's had to manage a team of individuals in such an environment, there are a few key responsibilities that I firmly believe are specific to a manager.

Keep them out of their comfort zone
In this type of environment, everyone needs to continually push themselves and each other. There is little room for corporate structure or maybes. It's about action and execution. To be able to function like this it is necessary that people are continually pushed outside the edges of their comfort zone. The goal is not to push them too far but just enough that it becomes a growing experience personally and professionally.

I've been out of my comfort zone for many years. The last time I found it, I quit my job.

Let them make mistakes
If you want your team to be effective you need to trust them, you need to let them run, and that means not being afraid of them making mistakes. Mistakes will be made regardless of how you proceed so why not try and set a positive spin on them. When mistakes happen, support them and move forward. I can assure you that the best learning and growth comes from making mistakes when outside of your comfort zone.

Keep them fed
Yes, keep them fed. Whether it is food (cheese, corn nuts and chocolate covered pretzels are very popular in our office), nice chairs, big monitors, an afternoon on team segways, friday wine-o-clock, or silly pictures on the walls, having some balance keeps us all grounded.

It is quite incredible that when you have good people together that good things happen.

Was the Old Spice Campaign Nothing More Than a One Night Stand?

noreply@blogger.com (Paul Neto) — Thu, 28 Apr 2011 15:17:00 +0000

The Old Spice viral social media videos were a well executed social media campaign and touted as a prime example on how an old brand can leverage social media channels to build awareness of it’s brand. While the total numbers of views, tweets and mentions are impressive by any measure (including 55% increase in body wash sales), one could argue that they failed in a primary social objective of maintaining the relationship. As great as it was, it simply came to an end. Thomas Baekdel goes as far as calling it a one night stand:

“They essentially created a one-night-stand. Huge amount of affection, but the next morning there is a note saying: “It was fun, but that was all it was.”

The thing with one-night-stands is that it can be really exciting, or a terrible experience afterwards. It all depends on people’s expectations going into it. Old Spice forgot to meet people’s expectations, and let many people down.”

P&G built a base of over 600,000 fans and the question remains whether they are really fans of P&G, Old Spice or Isaiah Mustafa? I’d like to see that metric explored. Hate to think that there are 600,000 people out there waiting for a call the next day?