Tag Archives: data

Amazon, The Data Company, Not The Retailer

Last week I blogged about $GOOG.  This week, we’ll talk about $AMZN.

I find myself using Amazon much more than I used to.  I use the desktop version and the iPhone app though most of my usage comes from the iPhone app.

I’ve tasked myself to think about why I use Amazon.

Price was my first reaction.

Buying online for consumers used to be about price – mainly because you did not have to pay sales tax.  Amazon is collecting sales tax, but from just 8 states but three of those states are the most populated states in the USA including California, New York, and Texas.  Based on a July 2012 estimate, they collect sales tax on 35.45% of the country.  New Jersey and Virginia are expected to join the Sales Tax list in 2013 which would push the total coverage of the USA to 41% of the US population.

So eventually, buying online will not be about saving on sales tax.  It’s about something much bigger.  Convenience.  Selection.  And much more.

Here’s what I basically netted out to:

Amazon isn’t necessarily about the lowest price, but the most convenient shopping (1-click).

The 1-click shopping experience on Amazon is amazing.  It’s dead simple.  I cannot tell you how many times I’ve been sitting with someone who recommends a book and I pull out my phone and within fifteen seconds, I order the book and its delivered via Amazon Prime (48 hours).  Their search works, their catalog of product is deep and broad, and the checkout experience is headache-less.

I wish every single site and store deployed a similar checkout mechanisms.

At the agency, we’ve partnered with Amazon in an advertising relationship and are working with them on a client business or two.  I never thought I’d be working with a retailer in this capacity, but I am.

So what is Amazon?

To me, in my opinion, is a data company that uses intent and e-commerce to build its dataset.  Sound similar to Google?  Yep.  Only difference is that Amazon is a bit more diverse with its data set as it has the actual sale of product.  Keep that in mind.

Oh, and Amazon also has an entire cloud hosting division.  How much data resides and passes thru their cloud?

 

 

Thinking About Google Fiber

I spent some time thinking about Google Fiber and it’s opportunity and threats to the consumer and business ecosystem.  This is not supposed to be a fully thought out piece but rather some raw thinking that I’m putting out to the Interwebs for continuing a conversation that was started by many of you.

Why is Google building out a Fiber Network?

After reading many articles on Seeking Alpha and other arm-chair analyst blogs, I believe the primary motivation to their planned (and current – Kansas City) fiber rollout is to protect their business and to allow for future growth.  What they are creating is stability for themselves and the pipes in which they can deliver whatever they’d like.

Lets think about this.

Google has done pretty well at maximizing paid search thru harvesting intent on the web.  This is a genius business and will continue to do well, even if there is an advertising downturn (lets hope not).  GOOG’s market cap of around $230 billion which is about 6x 2011 GOOG advertising revenue ($37.9b) could use some more growth prospects to satisfy investors.

Video is a big growth area for Google and YouTube will be its delivery mechanism.  YouTube streams over 4 billion videos per day and is looking to increase this (and the quality of videos) thru YouTube’s original programming partnerships.  The top 5 television markets are New York, Los Angeles, Chicago, Philadelphia and Dallas-Fort Worth; these account for roughly 12.5mm households and there is also some serious purchasing power in these markets.

You can bet that Google would like to put Fiber down in these markets as to gain access to distribute YouTube original programming distribution.  Note, they can currently distribute any of their content in market but if the last mile players like $TWC and $VZ utilize metered pricing, then YouTube videos might face consumption challenges.  If Google delivers the last mile, then viewership of YouTube content theoretically would be free as Google would be monetizing it from the other side (or at least Google would control it’s destiny).

Data is abundant and Google wants to learn from as much data exhaust as possible to make themselves better.  Larry and Sergey are engineers.  They are excited about technologies and want to build big systems.  Google Fiber will give them access to a truly big data set that they will then be able to tap into in order to optimize their systems for better content experiences and advertising delivery.  Advertising deliver is extremely important as Google derives much of its revenue from Madison Avenue.  The better the dataset that Google has, the better performance its advertising should yield.

Just thinking outloud about Google Fiber.  Did I miss any major points?

 

Search is bought, not sold

Over the past month or so since I first heard the line, “search is bought, media is sold,” I’ve been thinking about it as it’s resonated as both a media planner/buyer and a marketing/ad tech investor.

I went thru the exercise to understand this statement at its core:  why is search bought and why is media sold?

Search on any advertising campaign is pretty much a “must buy.”  Why?  Search is about Intent.  My colleague Taylor writes about it here and here.

Dictionary.com defines Intent the state of a person’s mind that directs his or her actions toward a specific object.

If a person is telling a search engine that they are already going to a specific object, then making it as a frictionless as possible to move said person from search to object is what a search engine delivers.  Pretty simple.

Search is almost cheating from an advertising perspective.  Yes, I said it.  There is no genius or persuasion to search.  Google built an empire that will keep returning cash until their search engine is no longer relevant.  The $200B market capitalization isn’t because Google has built a much better advertising technology, but because Google is playing on the key insight of Intent.  It was and continues to be genius; I wish I had done that.

While search ad copy is extremely important, especially on competitive keyword terms, capturing the Intent is pretty much a numbers game done thru the smarts of the search engine marketing (“SEM”) technologies and the SEM/biddable media analysts using them.

Media is sold, which means that a sales person (or API) needs to contact an agency or marketer directly, build a relationship, and sell in their wares.  This is very different than search.  I like to think that nothing is a must-buy other than paid search.

A good sales person with relationships and a great product must sell to be part of a plan and make a case why.  Alternatively, and increasingly, we’re using API’s to connect us to programmatic sources of buying non-search media that is removing the day to day sales person out of the picture.  Industries evolve.

Data to enhance targeting generally should win out on a media plan before any time of blind or proxy-audience buy.  The picture below illustrates this notion.  The old way of buying media was to start broad and then narrow down based on performance.  Now, it’s about starting narrow because we can, and then getting broader if it’s needed and budget allows.  It’s not that marketers are spending less, it’s that they are more refined in their targeting.

Targeting

What this overall thought proves is that the use of data, at least in the “intent” stage, will drive results.  It’s validated;  $200B worth of market capitalization, validated.

This leads me to an investment thesis of who else is capturing Intent beyond traditional search engines and how can we partner with those companies to accelerate them.  If you know of any, I’d love to chat.

Data Alone Is Not A Winning Proposition

I was fortunate enough to be asked by Ari & David Goldberg to speak at their State of Style Summit which was held today at the 92st Y in Tribeca.  They threw an A+ event and the turn-out of attendees was awesome; it looked like standing room only from the stage.  Job well done, Goldbergs.

On stage, I talked about data and the application of data for marketing along with Joe Zawadzki from MediaMath and Albert Azout of Sociocast.  I was on a tangent a bit and gave the crowd a laugh with the following quote:

It got re-tweeted a lot.

Looking back at it, it is immensely important.

Data is at all of our finger tips.  When you step on the scale each morning, look at your fitbit stats, log into your Mint.com account, or even review your Amex charges, you are looking at data that can then be turned into insights and then be actioned upon.

However, data alone does not mean action.  When I step on the scale in the mornings and am trending towards Alec Baldwin rather than Ryan Gosling, I’m not actioning data.   This is important.  Data alone does not make decisions.

An organization built for the next century is one who has to be able to wonk through large datasets, find insights and action them.  Just having data alone is not a winning proposition.  It’s the application of data, the extrapolation, and understanding that will lead to competitive differentiation.

If I was actioning the data from the scale, I’d not be eating this delicious chocolate chip cookie and tea from Mae Mae Cafe as I wrote this post.

The Bloomberg Advertising Terminal

(originally posted on Google Plus and then picked up on PaidContent)

I’ve been spending a lot of time thinking about data recently and it’s become the central investment thesis for kbs+p Ventures, our go to market approaches for The Media Kitchen, and how kbs+p communicates vision. My friend and entrepreneur extraordinaire +Jon Steinberg says it extremely well, advertising is becoming “guided by math, but moved by art.” For many folks in direct mail or other quantitatively driven marketing disciplines, this has been the norm, but I’m loving how this new norm is playing out across all of marketing.

Data isn’t new. Data allowed ancient salt traders to make important investment decisions in Egypt. Data allowed Christopher Columbus to accidently find the Americas. Data allowed Babe Ruth to know which pitches to throw to which batters. It’s been around.

Why it’s become a central thesis to us now is because it’s more actionable than ever because it’s become almost tangible and tools allow it to be ever more moldable. As a focus group of one, I use data to optimize my fantasy football teams thru +Nik Bonaddio‘s Numberfire platform, I use data through our Trading Desk, Varick Media Management, to optimize our biddable media campaigns, and I use data to help me understand where to invest my personal capital to help drive returns that can pay for my kids college tuition and my wife & I’s retirement.

As above, “data” can be used for many different uses.

One area of use that I’d love to see built out (and maybe I’ll pursue it) is legitimately, The Bloomberg Terminal for Advertising Data. If you are in the advertising technology ecosystem, then you’ve probably heard a million pitches with the words, “Bloomberg Terminal” but I think this is a huge opportunity around a very structured product. Let me explain.

Fact: hundreds of millions of dollars (if not billions) are being invested in media impressions thru biddable media sources

Fact: brands and agencies are building RTB advertising technologies to take advantage of market opportunities

Fact: publishers are going through an evolutionary period in which they transact their “wares”

Fact: agencies are in an evolutionary period in which they structure their buying decisions and put data front and center

In my theoretical world that I like to play out in my head every now and again, and run past trusted sources, I play out a scenario in which Advertising Traders have multiple screens on their desk, similar to a Bloomberg Terminal in which software is running showing the market dynamics and pricing. This Bloomberg For Advertising will show specific marketplace pricing (AdX, RMX, Rubicon, etc) indexes, demand volume, specific data asset pricing & demand (3rd party data), and the like.

To create this and carry out the vision, I believe as of now, but could be convinced, that this needs to be executed by a unbiased 3rd party company who isn’t tied to media or data volume. They purely are (profitably) motivated thru licensing of their Bloomberg for Advertiser software.

Why is this important?

1. Data assets as simply described above are going to become increasingly important for investment decisions in the near and mid-term.

2. Publishers need access to this information the exact same as advertisers to help drive their businesses forward.

3. Regulation of markets is a commonplace in the USA and the advertising marketplace is heading in this direction, at least at a preliminary level, especially as we increase our usage of spot and forward markets.

I believe there is a very large opportunity to be this for marketing and advertising. If you are out there building this or have a viewpoint similar or dissimilar, I’d absolutely love to hear from you.

Measuring Integrated Advertising

I have a lot of respect for Mark Suster, an entrepreneur who turned venture capitalist and now is investing out of GRP Partners.  He writes a terrific blog called Both Sides of the Table and his posts are picked up on TechCrunch and other major outlets.

He recently wrote a declarative post called The Future of Advertising Will be Integrated.  The post went up on April 29th but I’ve been noodling it ever since. Due to some personal obligations, I’ve not been able to respond, but finally, here it is.

As an entrepreneur turned ad agency guy, when I hear the word “integrated,” I immediately think media and creative under one roof, such as my firm, kirshenbaum bond senecal + partners.  I personally believe this is really the only way to go if you want to get to a big platform.  With media and creative all under one roof, under one P&L, and with a cohesive team, you can create big ideas that know no creative or media boundaries.

We have a saying internally at the agency, E=(MC)2, which is obviously repurposed, but it means a “[brand] experience” is exponentially greater when media and creative work together.

Enough agency speak for now but keep this last sentence in the back of your mind as you read the rest.  I hope you do.

Mark came at his post a bit differently and took the above tenants, whether he realized or not, and applied them to the digital media ecosystem today.  He highlighted a few companies such as my buddy Ari’s company, Solve Media, along with Adly and Kontera (amongst others).  The creative is the media in most of these, along with the media being the creative.  I’d argue Paid Search links play here as well.

The Elephant in the Room

One of the largest issues that the digital advertising ecosystem faces today is that we as an entire industry, are not setup to measure the “integrated” nature effectively. Because of this, at scale, this is not a near term reality.  There, I said it.  The elephant is in the room.

The digital advertising ecosystem by default rewards the intent harvesters, not the intent generators.  The primary reason why is that many agencies and marketers are using 3rd party ad serving systems that reward the last click or last action.  In the world of rewarding the last click or action, generally, the ad networks are the ones who win out.  There are 400 (or 700 depending on who you talk to) or so ad networks in the world who have nice businesses.  Just look at ValueClick or InterClick’s financial statements as they are public.  Not too bad.

THE Digital Opportunity

Because of the above, therein lies an opportunity.  If we believe what Mark wrote last week and I’ve been saying for years, then an opportunity lies in being able to create a measurement platform that allows us to understand intent harvesting and intent generation/creation. Piecing together a DART (3rd party ad server) report with a ComScore or Knowledge Networks study is inefficient and frankly, annoying.   There needs to be an evolution here.  This is a big opportunity.

Where We Are Today

Many readers of this blog don’t work in advertising agencies but are awesome entrepreneurs looking to figure out the next big idea to go and tackle.  Being that you are not in the walls of agencies on the daily basis, I thought I’d take the remainder of this post to outline where the industry is in terms of advanced analytics and then open this up for commenting in the thread below.

I highly request that you engage in the comments as group knowledge will benefit the community at large, you might find your next co-founder, and I love open conversations.

Ad Serving:  The Madison Avenue ecosystem basically uses one of three third party ad servers to “serve” and “track” different pieces of creative.  We use Microsoft’s Atlas, DoubleClick’s DART, and MediaMind.  In Q1 2011, we moved the majority of our clients off of Atlas and onto MediaMind because I personally have a strong viewpoint of independence of my ad-server and it’s relationship to media. (should be separate)

Data Warehousing:  This is a relatively new area and somewhat unchartered territory for many agencies.  Many agencies rely on their third party ad-server to be their main data warehouse for tracking. This is good, as you’d be surprised how many people don’t use a 3rd party ad server, but this is not great. Using a full on data warehouse such as VisualIQ, Neteeza, Artemis, or others allows for a larger capability to manipulate data and understand the relationships between touchpoints beyond “last click.”

At the agency, we’ve been using VisualIQ with some of our most progressive clients and the reports and results we’re seeing are fascinating.  One of the biggest questions we’re tackling is “optimal touchpoint analysis” and we’re seeing the relationships between display, video, search, social, and beyond.  We can now determine a value to each one.

Brand Lift Studies:  While I’ve argued time and time again, that “brand” advertising for the sake of brand advertising online is dead, many marketers continue just spending on “brand.”  Agencies use 3rd party brand study vendors such as ComScore, Knowledge Networks, Vizu, and others that help measure the “lift” (or change) associated in any one of many categories including but not limited to awareness, intent, and consideration.

Opportunities

·     The basic ideas behind today’s ad serving systems were conceptualized in the mid to late 1990s.  Online video, social, search, etc were not around then.

·     Product placement and integration into online video and social are hard to quantitatively measure with a 3rd party ad serving system as the only metrics you can pull back to your ad server are by using a click-tag.

·     The Display ecosystem is being fractured into traditional display (i.e. banners on ESPN) and social display (i.e. creative/textual units on Facebook, LinkedIn, etc).

·     I see Paid Search and Display converging on each other within the next 12 months. In some cases, they already are: Google Content Network.

·     I challenge you to ask your 3rd party ad-serving vendor to recommend an attribution model – report back what they tell you.  Not much – there is no standard yet.  It’s unchartered territory.

Next Steps

I would obviously love to hear your feedback.  Please post it in the comments section below or shoot me a note.  I believe that we won’t see large integrated opportunities that get their portion of the measurement/attribution credit until there is a way to measure these.  While we might try one or two of these integrated opportunities on each media plan, if you ask the agency how they really performed, the agency won’t have much to tell you because the tools for measurement are ancient.  With the data warehouses mentioned above, we get much better, but not perfect.

While we don’t need perfect to make the industry move forward, we do need better tools.  If you are building them, I’d like to speak to you.

Data Dashboard: eBay Auctions, Autotrader, Manheim Data Sets

I’m in the market to make a new automotive purchase. Because of the nature of the purchase, I’d like to do as much due diligence as possible.

I’ve spent the last week watching new auctions pop up on eBay Motors and the like for a few specific models of cars, but realized that unless I keep a detailed manual spreadsheet of all I’m tracking, I’m going to be wasting my time watching different auctions.

The data geek in me is excited.

I’m wondering if there are any data dashboards or software that can scan eBay, Autotrader, Manheim, and other automotive datasets and record the following:

  1. Automotive Make (i.e. Toyota)
  2. Automotive Model (i.e. Camry LE)
  3. Exterior Color
  4. Interior Color
  5. Mileage
  6. Condition
  7. Source (eBay, Autotrader, Manheim, etc)
  8. Seller Name
  9. Seller Email
  10. Seller Phone

If I had just the 10 items above, in a structured data set for a rolling 90 days of data, I think I’d be very much ahead of the game.  I’d then port the data table into visualization software such as Tableau and then manipulate it so that I can figure out which cars I should be focusing on.

So, my question to the Interweb:

  1. Does this type of software already exist?
  2. If not, can it exist?  (if so, please contact me if you’d like to build it)

I don’t think this is too difficult but I could be wrong.

Ari Gold & Moneyball Theory

I’ve always been fascinated with how talent agencies operate.  Back 4 years ago, I wrote about creating a Talent Agency for video game athletes. The idea of a “Talent Agent” is sexy to me because they get to identify and grow people.  In some ways, I do this for early stage technology, as both an investor and advertising agency executive.

While I’ve not walked the halls of Creative Artists Agency or William Morris Agency in quite some time, though I did back in 2001-2003 when I had a music company, I would hope that they are investing in tools and capabilities to digitally mine petabytes of data to identify the next Justin Bieber or Lady Gaga.  I’d argue it’s all in the data, similar to Moneyball, and finding the next big franchise requires servers, algorithms, formulas, and a few quants.

The story of Bieber is told almost every day now – his mother uploaded videos of him to YouTube playing various instruments and they started gaining momentum… and was found by a music industry exec and ultimately signed to a big deal. Now he’s a huge pop star sensation and a true early stage success story.

But what interests me is less about the star herself, but the tools and resources necessary to identify the potential star.

Lets start with YouTube since we spoke about it above and it’s the lowest hanging fruit.  With billions of views and hundreds of millions of viewers, the data that YouTube has on video viewership is tremendous.  Are there any companies who have built platforms on top of YouTube that allow the big (or small) talent houses to be alerted when videos are catching fire (or going “viral”) or the ability to track certain upcoming artists for their views and ratings?  If CAA or WMA wasn’t doing this, it’s almost negligent.

Could CAA or WMA be one of the largest purchaser or subscriber of the bit.ly database?   I’d hope so.

But digital doesn’t just start and stop with YouTube.  What if you could take all the YouTube output from above and cross reference that with Google Insights for Search: and examine all the trend volume around specific artists or videos.  By doing this, you get a linguistic and geographic context to the output.   Is Artist A trending in Sweden and Switzerland or just one of the above places?

Taking YouTube and adding Google Insights for Search is getting better, but what if you could layer a social media monitoring solution on top of this?  So for an A&R executive at a record label, or an artist manager at a management firm, they could not only see media hits (stories), but sentiment and momentum numbers?    There are way too many monitoring solutions in the marketplace today so I’d have to think that one or two of them have cracked the code to sell into the talent industry.

Trendrr has done some interesting things for this space as they allow the pivoting of different datasets together.  You can pivot Gaga album sales (Amazon) with Twitter followers to see if there are correlations.  A good start, especially in bringing together different databases and making them queryable.

While I’ve done virtually no research into this post about whether or not these platform(s) exist or not, it’s something that should and someone should build it.

Transparent Data Streams

I’ve been wrestling with transparency lately – mostly around the consumer side of things.  Transparency is great, but if it comes at the cost of too data overload, then is it really worth it?

One of my blog readers reached out and said he’d started a service called VoyURL of which I’ve been participating with for the past 24-36 hours or so.  It’s interesting – right now as a novelty, but once I can use all of the data to extract meaning, it could have implications.  I’m not quite sure what they are yet.

I’ve pulled some initial reports of the entire VoyURL universe and my own.  It’s fun for me to see how my web consumption differs from the VoyURL universe.

It’s like Blippy for browsing data.

VoyURL Data

Enhanced by Zemanta

Opportunities in Advertising Ecosystem

I gave a presentation at the First Round Capital CEO Summit yesterday in San Francisco on the subject of areas that need innovation within the advertising ecosystem.  While this is a very very broad topic and I only had 45 mins or so, I had to narrow it down to a few areas to speak in-depth about.

I modified the presentation for public consumption and reformatted it from it’s original version.  The presentation is obviously a “talking” presentation with much of the context and clarity coming from my voice overs.  I tried to put a few sentences under each topic so you can get a sense of what each slide is about.  It’s short and sweet – and would love to hear from anyone who has any “solves” for these opportunities.

Here is the link to the presentation on Scribd

Opportunities in Online Advertising

I had a great time at the presentation and met some awesome entrepreneurs and CEOs.

Enhanced by Zemanta