Tuesday, August 25, 2009

Photography Series: All That Glitters Is Not Gold.


From composing words, to composing images.


The theme here is the contrast between a meditative stone buddha and a glittering gold column behind it.


Taken with a Sony DSC T1 digital camera. Taken in the afternoon sun, while lying on searingly hot stone to get the right light for this photo.


Location: Temple of The Emerald Buddha

Wednesday, August 19, 2009

Data, Privacy and Social Media: Impact on Analytics and Tools

Executive Summary: How does data, along with the constraints on it, impact web analytics tools and the industry? How do you start structuring your thoughts about data's role in this industry?
Note: This came about after some rumination on the way the web and social media are being used by businesses, and this post has been updated with some of those thoughts to help continuity.

My last couple posts have been threatening to run down this alley for a while, but I realized that Data, Privacy and Social Media need to be addressed together in a separate post. I shared these thoughts with Alan Chapell, Chapell & Associates, after a chance meeting with him at an event before the Affiliate Marketing Summit.

Do Privacy Issues Impact Data Access for Analysis?
Yes, especially when you are not the primary service provider.

When you think about Data, Privacy and Social Media in the same breath, issues like those that Facebook has recently faced come to mind:
1> The criticism:
http://consumerist.com/5150175/facebooks-new-terms-of-service-we-can-do-anything-we-want-with-your-content-forever
2> The response:
http://blog.facebook.com/blog.php?post=54434097130

It is important to note that this recent case is part of a continuum of controversies over access to and sharing of online customer information during the past decade.

Data privacy is just one of the factors impacting the web analytics industry. Hence, to assess the impact of data related privacy issues on web analytics for social media, you need to drill down for insights in the following manner:

Drill Down Level 1: Analytics, Market Structure and Client Trends
1> State of the Art in Analytics.
2> Client Requirements.
3> Market Structure.

Drill Down level 2: Client Needs and Impact of Data
Insights from areas 1, 2 and 3 above will help you create a picture of:
1> Industry Objectives: Met and Unmet
These are objectives that the web analytics industry can currently help its clients meet. E.g. What metrics can be measured, and how do they translate into impact for the client?

2> Data Access and Usage: Reality, Needs, Gaps
This is role data currently plays, and can play, in the industry. E.g. What data, out of the available data sources, is currently utilized, how is it utilized, and how can it be utilized?

Drill Down Level 3: Data Privacy Constraints
Once we have this level of analysis, we can gain an insight into the impact of current privacy constraints on the industry. Importantly, we also can overlay various combinations of data privacy constraints to get an insight into the role data privacy can play in the industry.

Drill Down Level 1: Below are some questions that help us get insights:
1. State of the Analytics Art: What is the level of skill and tool maturity of social media companies in terms of tracking social networks, and allowing ad targeting around them? Some additional evaluative questions can be found at my previous post below:
http://randomjunkyramblings.blogspot.com/2009/08/ads-applying-web-analytics-tools.html

2. Comparing Traditional and Social Media Analytics:
An alternative to consumer targeting via social network analysis is targeting via demographic and psychographic analysis. Here companies like Nielsen have developed a set of consumer personality profiles. Are social media sites developing such profiles?

3. Third Party Analytics:
How do third party analytics compare with the State of the Analytics Art? Specifically, (in terms of providing social network access to third party providers for analytics) are social media companies taking an API and metrics sharing or a raw data sharing approach? This impacts the competitive environment in the analytics industry.

4. Client Needs:
What are clients doing in the web and social media space? What are the client objectives in this space? Are they looking at demand generation or demand fulfilment? What would they like to/ need to measure?

5. Market Structure (and Opportunity!): Who are the major players? Is the market fragmented? Are there alliances in this market? How strong are service provider- client relationships? Do analysis, metrics, or data clearinghouses exist?

The opportunity!: Is there a market for "clearinghouses" for analytics that would provide media buyers access to social media consumer data across platforms? In reality, a "clearinghouse" is not easy to achieve- E.g. In the brick-and-mortar retail world, Walmart does not share its scanner data.

The Background
How did I arrive at this post? My last couple of posts have been pointing toward looking at Data, Analytics and Social Networks through the privacy lens:
1. Data, Privacy and Customized Brand Metrics:
http://randomjunkyramblings.blogspot.com/2009/08/sustaining-brand-conversation-behavior.html
How do you deal with a profusion of metrics which often have an unclear context? We need metrics with clear semantics. Some of these metrics may be custom created for a specific brand, consumer profile, activity and social media context.
Future Shock
While generic, industry standard metrics are important, there is a huge *future* potential for metrics customized to the brand. These will hinge on the data collection capabilities of social media platforms, and their ability to share it in a cheap, safe, anonymized manner with third parties for further analysis. Additional factors that come into play- quality of data, privacy concerns and analytics capabilities.

2. Data, Privacy and Web Analytics Tools:
http://randomjunkyramblings.blogspot.com/2009/08/ads-applying-web-analytics-tools.html
3. Can you utilize the social network based targeting tools in conjunction with non social network based predictive analysis tools to minimize errors or optimize marketing spend?
Don't Forget Data: The Root of All Analytics
While it is easy to get lost in the tools and their trade-offs, don't forget the data. Your data- what attributes are available, access, usability and quality- may dictate the choice of tools.

What do you think?


--

Web Analytics: Social Network Analysis Tools for Ad Targeting.

Executive Summary: Are you looking at social network based ad serving analytics tools? How do they stack up on the measures below?:
1. Are they mature enough to provide "error" analysis (say Type I/ Type II) on the targeting?
2. Can you compare the "error" from social network based analytics tools against the error using non social network based analytics tools?
3. Can you utilize the social network based targeting tools in conjunction with non social network based predictive analysis tools to minimize errors or optimize marketing spend?

Why is there no focus on cost above? Apply cost factor weights to the results from the questions above and you have your ROI!

How To Develop Criteria For Evaluating Social Network Analysis Tools
From the Blueprint-for-winning-Social-Network-Analytics-Orders Dept., the How-Do-I-Start-a-Competitive-Analysis-of-the-Marketplace desk.

What About Social Network/ Graph Analysis for Marketing?
I have tinkered with various "traditional" analytical tools, like applying predictive modeling for web analytics, to looking at web server log files to analyze "eyeball" trends during the early days of the Internet boom, so, a recent post below reignited my interest in social network analysis for serving targeted ads:
http://www.marketersstudio.com/2009/08/the-social-graph-ad-targeting-buyers-guide.html

The post referred to a Knowledge@Wharton article on Network Based Marketing:
http://knowledge.wharton.upenn.edu/article.cfm?articleid=1637

This article had some interesting insights:
Insight 1: Social Networks Help Us Find Like Minded Buyers
Another possibility is that because people often tend to talk to people like themselves, their buying tastes would be similar regardless of whether they ever discuss the product. "Social theory tells us that people who communicate with each other are more likely to be similar to each other, a concept called homophily," the researchers point out. "...Linked consumers probably are like-minded, and like-minded consumers tend to buy the same products."

Insight 2: Social Networks Help Target Buyers More Effectively
"In addition, analyzing the network allows the firm to acquire new customers who otherwise would have fallen through the cracks, because they would not have been identified based on traditional attributes."

The Power of Non Social Network Analysis Tools
I can attest to the fact that with non network based- a.k.a. traditional analytics- tools provide you incredible ways of reaching out to your target audience. If you have the data- various demographic, psychographic or geographic attributes- some of these tools can help you:
1. Target your audience effectively,
2. Give you a sense of the targeting errors (Type I/ Type II) with the model,
3. Help you rank order your targeting efforts by efficacy, and hence optimize your marketing dollar.

Comparing Social Network Based Marketing Tools vs. Existing, Non Social Network Tools
The first thought at comes to mind is whether social networking based ad targeting tools have matured to provide these features. Specifically:
1. Are they mature enough to provide "error" analysis (say Type I/ Type II) on the targeting?
2. Can you compare their "error" against the error using non social network based analytics tools?
3. Can you utilize the social network based targeting tools in conjunction with non social network based predictive analysis tools to minimize errors or optimize marketing spend?

Don't Forget Data: The Root of All Analytics
While it is easy to get lost in the tools and their trade-offs, don't forget the data. Your data- what attributes are available, access, usability and quality- may dictate the choice of tools.

In the Future Shock section of my previous post below:
http://randomjunkyramblings.blogspot.com/2009/08/sustaining-brand-conversation-behavior.html
The ability to develop customized brand metrics "will hinge on the data collection capabilities of social media platforms, and their ability to share it in a cheap, safe, anonymized manner with third parties for further analysis. Additional factors that come into play- quality of data, privacy concerns and analytics capabilities."

Data privacy currently remains an important factor- more in a later post.

Aren't You Forgetting Computing Power?
I am discounting computing power as a factor. I will not tell you why. :-)

Conclusions For Social Network Based Ad Targeting
Non social network based marketing tools are pretty powerful. The questions above give you some insight into decisions related with integrating your existing analytics capabilities with social network based analytics. For the service providers, this is your blueprint for winning orders.

What do you think?

--

Monday, August 10, 2009

Sustaining a Brand Conversation: Behavior Tracking and Measurements Notes for Brand Marketers

Executive Summary: How do you deal with a profusion of social media metrics which often have an unclear context? We need metrics with clear semantics. Some of these metrics may be custom created for a specific brand, consumer profile, activity and social media context. We could create two categories of metrics- generic, infrastructure metrics of the type that are commonly thrown about and need contextual understanding, and functional metrics that have clear semantics attached.

The Story So Far
We covered a need for strategic thought behind social media marketing investments below:http://randomjunkyramblings.blogspot.com/2009/07/to-strategize-or-not-to-strategize.html

The next step is to generate a picture or customer touchpoints/ interactions with the brand across various channels. Besides qualitative insights, you would like concrete measures that support these insights.

There are challenges in tying in consumer behavior in a "regular" distribution channel with that across social media channels. Leveraging your existing, real world consumer profiles in the social media world is a separate theme. Our focus in this post is to find ways to measure and track consumer behavior in the social media channels.

Challenges with Interactive Metrics Today
There are two challenges with social media metrics today:
1> A profusion of metrics.
2> A need to understand the context in which these metrics are being generated.

David Berkowitz has a great post here on the various metrics available to marketers today, and a proposed Cost Per Social Action (CPSA) metric:http://www.marketersstudio.com/2009/08/cpsa-cost-per-social-action-the-new-pricing-model-for-social-media.htmlThere are third party companies like Visible Measures doing interesting work as well.

As for context, a wise man once said, context is everything. Does a metric mean the same coming from a face to face interaction as opposed to one over twitter, or even one from a different social media platform?

A Potential Solution
Some new metrics are needed. However, they need to be functional in nature. By functional- I mean that the metrics need to carry a consistent meaning for brand marketers. CPM clicks could be meaningless in some contexts. You might argue that this is true of all metrics. True. Hence the need for metrics with specific meaning and context attached to them.

I am not saying this is the end of the existing metrics. We could have two classes of metrics- the infrastructure metrics and the functional metrics. All metrics have semantics, hence I am calling these new metrics "functional" metrics, instead of calling them semantic metrics.

Future Shock
There lies the key. While generic, industry standard metrics are important, there is a huge *future* potential for metrics customized to the brand. These will hinge on the data collection capabilities of social media platforms, and their ability to share it in a cheap, safe, anonymized manner with third parties for further analysis. Additional factors that come into play- quality of data, privacy concerns and analytics capabilities.

What Can We Do Now?
While it is great to theorize about the future, there is opportunity today to develop measures that make sense for a specific brand, consumer profile, activity and social media context.
You are welcome to contact me for a conversation on these.

Update 1 (Thanks to David's followup): Functional Metrics Example
Craig has a great illustration of my distinction between infrastructure and functional metrics here:
http://www.funnelholic.com/2009/03/12/memo-to-the-cfo-3-lead-generation-metrics-that-matter/
Cost Per Lead (CPL) could be called an "infrastructure" metric, as opposed to Cost Per Opportunity (CPO) which could be called a "functional" metric. CPO is tied to the lead and pipeline generation funnel, and not to the various tools and mechanics that cause CPL number variation. CPL feeds into CPO generation.

Update 2: Caveats and Another Functional Metrics Example
The challenges the metrics are expected to address:
1. In metrics, we often miss the forest for the trees. As I have mentioned before- Marketing is following Social Media.
2. Given the context of the million dollar Superbowl ads, we need to build the kind of "bridges" in social media that already exist in traditional media and which allow traditional media to justify its spend to some extent. That's a separate problem.

So taking the sales theme further (you can see I am trying to leverage my B2B sales/ account management experience) here's what I would call an infrastructure metric derived out of a sales force effectiveness ratio: social media effectiveness ratio = social media "wins"/ customer "contacts".

Now, you might call sales force effectiveness metric a functional metric that has been translated into an infrastructure metric. True, wins and contacts are tied to the platform. We then build a cross platform metric that takes this data and spits out the "functional metric" results.

The Four Philips brand equity measures- Uniqueness, Relevance, Attractiveness and Credibility- are a tougher portability nut to crack. However, a quick metric that is "translatable" that would be familiar to brand and category managers- ACV.


What do you think?


Additional Background
A backgrounder to help you develop your own perspective on the ideas here:
http://randomjunkyramblings.blogspot.com/2009/06/brands-economics-blink-twitter-facebook.html


--

Sunday, August 9, 2009

VC Decision-making: Investing in an Existing Portfolio of Companies.

Executive Summary: Investing in existing portfolio companies involves a lot of groundwork to be done by the Venture Capital General Partners. Below is an experience recounted by a VC investor, followed by some of my thoughts on the tale well told.

The VC Doubling Down Experience
Fred Wilson has a great post on "doubling down" vis-a-vis portfolio companies:
http://www.avc.com/a_vc/2009/08/doubling-down.html

The Analytical Backgrounders
If you have read a previous generic post linked below, you will definitely enjoy an experienced VC investor sharing his experience above in making the decisions:
1. Creating Value through Operational Improvements: http://randomjunkyramblings.blogspot.com/2009/04/panel-creating-value-through.html
2. Microeconomics, Synergies and Operational Portfolio: http://randomjunkyramblings.blogspot.com/2009/05/microeconomics-synergies-and.html

Identifying a subset of "core"/ "sustainable" (your mileage may vary) investments from a portfolio during tough times is not an easy process even for great investors. The key aspects of Fred's post are:
1>raise funds during tough times, and,
2> choose to pump in more money into the newly identified "core" investments, instead of making new investments.
3> "restructure" existing investments in some way- e.g. strategy, team, cost.

To better understand the VC decision making process, here are some questions you may ask:

1> Fundraising For An Existing Portfolio
- Would you have your "core" investments list, and their follow up round funding needs in hand, when you raised funds?
- Would your fundraising account for possible iterations to this list? E.g. Would your fundraising process include discussions on potential exits and the potential returns to LPs from these exits?

2> Portfolio Picking as Cherry Picking
- Portfolio Company Consultations: Would you choose which investments needed more money from you on a case by case basis? Would this process be any different from board level discussions on company performance, except for more stringent criteria being thrown into the mix?
- Follow on funding Factors: Would the lack of (or unfavorable term sheet conditions in) follow up rounds of funding due to economic conditions force you into considering more investments in the same companies than you normally would?
- Finance Portfolio Constraints: Would you exit firms that had strong potential, but would skew the risk-reward profile of your portfolio?
- Exit Negotiations: What factors would cause you to take a "non core" company off the "for sale" list? i.e. What could cause you to exit the exit negotiations? How many "non core" companies would you have in play for exits at any point of time? Would you keep the window of being part of a follow on syndicated lending team on some "non core" companies?

Another way to quantify this is as follows:
- Company Funding Needs: How may of the newly identified core/ sustainable investment could do without funds from the same VC firm?
- Finance Portfolio Constraints: How many of the investments outside the "core" list had funding needs that the VC firm could not accommodate in the risk-return profile of the reconfigured portfolio, despite strong potential of meeting expected returns?

The Investor Basics
1. (Re)Evaluating financial options can be difficult in tough times over investments you have made which count on an expected future value, but that's par for the course for any investor.
2. The next step is to identify core investments that continue to be strong on performance indicators,and which, in your judgement, will provide strong returns. This too is par for the course for any good investor.

For those who have doubled down in tough times, a simple question to ask would be how many companies you exited went on to provide "spectacular" returns after their exit?

Its NOT So Easy
This only goes on to illustrate that a VC's decision making to exit portfolio ventures is, for starters, pretty difficult on the VC as well.

What do you think?


--

Thursday, August 6, 2009

The Startup Thought Process Series: Commuter Rants

Series Initation Note: This kicks off a series on my conversations with folks starting digital media companies with the objective of assisting the enterpreneurs.

The Startup: I met an owner close to rolling out a site for commuter rants. Simply put, the site is a forum for commuters to rant about their commute.

Our relatively brief conversation focused on helping him with the vision/ raison d’ĂȘtre for the site. Snippets of the conversation are listed below. These are really interconnected factors, however you need to be able to think through them linearly once, before you iterate through the options and interdependencies. As with a few startups, these answers may change with time, however, it helps to have concrete thoughts about these questions at the start of the journey.

I. Market Potential
The entrepreneur's first area of uncertainty was: how frequently would a commuter rant at his site? We broke that down into market sizing and frequency of usage.

i. Who is the target user of this site? What is the market size?
What kind of commuter? Someone who takes the NJ Transit or Metro North to and fro work? Or does it include someone stuck on the D.C. beltway on a Friday evening? The idea here is to understand an existing unmet need and customer behavior tied to this unmet need.

For sizing, there are several ways to generate the numbers- by geography, by demographics, etc.

ii. What would the growth and usage trends be like?
Would they be like that of Twitter (where 30% of the users tweet once never to return) or like that of Facebook?

II. Business Model and Market Strategy
We are really thinking about distribution channels, partners, customer relationships, core capabilities and revenue models here, all of which can be expressed pithily as:
Would you prefer a B2C model or would you modify the site for a B2B model?

Note: We explicitly kept aside market defensibility to assist in brainstorming.

A> B2C Model
i. How would you grow the B2C site?
Would you eventually develop features tied to hyperlocal search to enable customer stick? E.g. Regulars in a train compartment can connect with each other?

ii. How would you monetize the site?
Through Ads, and possibly, viral content (to help folks cool down, for starters)?

B> B2B Model
i. How would you grow the B2B site?
After an initial push to bring on site users, would you consider tying up with media companies who may leverage feed from this site? E.g. TV Weather and traffic update has a ticker running at the bottom which shows "selected"/ "near real time" commuter "rants"?

ii. How would you monetize the B2B site?
How many media companies would buy into this? What would such features be worth to the media companies?

III. Product Strategy
Would you roll this out as an independent site/ platform? Or,
Would you leverage existing platforms like the iPhone and/ or Facebook?

The questions for you:
1> How would you have looked at this differently?
2> Would you invest?
3> What changes, if any, would change your investment decision?


What do you think?

--