Product Strategy as Segmented Outcomes

Readers of the Innovator’s Solution should be familiar with outcome-based segmentation as a key technique to identify customers for a breakthrough product.  The essence of the technique is to think through “outcomes” that existing or potential customers want to derive from a product or a service, and tailor existing or new offerings to align with those outcomes. In simple terms, outcomes map directly to  customers’ value proposition for your product.

I think this technique should be an essential part of a product manager’s toolkit when you:

• plan your roadmap
• think of product extensions
• assess build vs buy opportunities
• map the competitive landscape
• identify segments to NOT focus on
• figure out positioning, messaging and value proposition for your products

An Illustrative Example

Imagine you have a product that allows non-technical users to build websites quickly. While such products have been around for a long time for building static/”brochure-ware” sites, you want to assess underserved/unaddressed opportunities. In particular, you believe online marketers might be an opportunity worth exploring further. The table below depicts the types of outcomes online marketers would be interested in, along with features that would support those outcomes.

	Target Market: Online Marketers
	SEGMENT: “Converters”	SEGMENT: “Conversationists”	SEGMENT: “Product Launchers”
Outcomes desired by segment	Drive conversions for an existing product or service through a website. Simply put, this segment values demand generation.	Allow brand managers to converse with brand enthusiasts and learn about brand strengths, weaknesses, threats and opportunities via a web destination. Simply put, this segment values engagement with consumers.	Create buzz for a brand new product or service through a website. Simply put, this segment values awareness.
Top Level Features (Examples)	Landing page templates based on conversion best practices SEO features A/B testing Support online forms for registration, surveys etc Tracking visits and conversions	Support various forms of user expressions such as video/photo uploads, comments, ratings, reviews etc Tracking user engagement (e.g. number of user-contributed videos, comments, reviews etc)	Support for premium/celebrity videos, high quality images “Virality” features (e.g sharing, syndication) Tracking overall awareness of product including page views, number of times videos/images are shared etc

The most important takeaway from this table is the value proposition that different segments derive from your product. For “converters”, it is all about creating a website that drives revenues. For “conversationists”, the value is around providing a destination for the brand manager to engage with brand loyalists. For “product launchers”, the value is primarily around generating awareness for a new product. You can use this table to drive the roadmap for your product. For example, you might want to focus first on the “converter” segment. You might decide to partner for the tracking/web-analytics feature which is common to all the segments. Similarly, you would want to map out the competitive landscape to this segmentation.

The Crying Game

I am delighted to announce a new technique to stop kids from crying when they decide to throw tantrums. Results are guaranteed based on one data point. Here’s the approach.

When the kid cries, tell him/her you are going to record it. I found my cell phone’s memo feature really handy to record crying spells. Once the recording is complete, play it back to the kid. He/she will probably be sufficiently surprised and distracted by the recording to stop crying. Some self-aware kids, like my 4 year old, might actually feel differently about crying when they hear themselves.

I quizzed my 4 year old recently about the experiment. She told me that she usuall “cries for nothing”, and was “embarassed” by the recording. We mutually agreed that such recordings will not be made public in her pre-school.

Convergence in Content, Commerce, Campaign Optimization

One of the things I am beginning to notice is the convergence in optimization technologies for ad targeting, content recommendations and product recommendations. In a prior article , I had talked about Context, Intent and Affinity as three techniques used to target ads on web pages and search results pages.

It turns out a number of products are beginning to use some combination of these approaches for content recommendations on content-oriented sites and product recommendations on eCommerce sites. In some cases, the same product can be applied to both content and product recommendations- Baynote being a prime example.

The table below summarizes what is out there. As you can see, these products use one or more of the following approaches to target content and products:

• context: use “aboutness” of content to recommend related content. Example: Sphere
• affinity/behavior: use click behavior of a given user or a cohort of users to cluster content and make recommendations.  Example: Loomia, Baynote, Aggregate Knowledge etc.
• popularity: use current popularity of stories to recommend content. Example: Techmeme, One Spot, One Riot.

Some untapped opportunities:

• Using twitter and facebook status to recommend content/products
• Using tweats to gauge break out topics. Would nicely augment products like techmeme.
• Providing video and image recommendations.

Product	Focus	Approach	Business Model
Aggregate Knowledge	Content Recommendations based on click-behavior	Works very similar to Amazon product recommendations. E.g. people who clicked on this story A also clicked on Story B, C, D.
Baynote	Content and Product Recommendations based on click-behavior	Works very similar to Amazon product recommendations. E.g. people who clicked on this story A also clicked on Story B, C, D.
Loomia	Content Recommendation based on click-behavior	Works very similar to Amazon product recommendations. E.g. people who clicked on this story A also clicked on Story B, C, D. Also offers video recommendations.
One Riot	Search results driven by real-time popularity of items	Collects data from browser beacons to gauge popularity of stories and use it to rank stories
One Spot	Content Recommendation- based on story popularity	Uses linking behavior between sites to rank and recommend popular stories on a given topic. Works similar to Google Page Rank.	$150/month per topic
Outbrain	Content Recommendations
Sphere	Content recommendations based on topic
Taboola	Video recommendations
Techmeme	Content Recommendation for Technology topics- based on story popularity	Seems to be surface most popular articles on technology at a given time. I strongly suspect that it uses linking behavior between sites to guess popularity of a given story.	Sponsored links priced at $5 CPM (actual pricing is several thousand dollars per month which translates to about $5 CPM at current page views)

A Tot’s Thoughts

For a long time, I have been very curious about what my child is thinking, from moment to moment. I think I have found a way into her thoughts, so to speak, using the following game. I told that I would count 1-2-3, then stop, and ask her for her thoughts. I would just listen to what she said, and move on. Then we would talk for sometime on whatever other topics came to us. Then I would count 1-2-3-stop and ask for her thoughts. Make a note and move on. And on and on.

I found it remarkable that:

1) she actually understood the game, and played along
2) I could actual glean what she thinks of beneath the surface, even as we talked about other things. In our game, for example, two of her thoughts had to do with her mother. How her mother had promised to drop her in school but did not. How she really liked what her mother had packed in her lunch for the day. Most of her other thoughts were random- she thought the ceiling fan was broken, something about Thomas-the-Tank-Engine that I did not quite understand and so on.

Incidentally, this is just another mindfulness technique, but applied as an external observer versus one’s own thoughts.

Competitive Positioning in Internet Advertising

Part 3 of US Internet Advertising Landscape

Read Part 1 here

Read Part 2

If you are a publisher or an advertiser today, there is a bewildering array of choices for you regarding how you monetize your content as a publisher or reach a desired audience as an advertiser. The final article in this series is an attempt to clarify how the different players fit in with respect to your objectives.

Refer to the accompanying figure as you read further.

Positioning in the Internet Advertising Landscape

Context, Intent and Affinity…the “CIA” of Internet Advertising

This is Part 2 of the US Internet Advertising Landscape. Read Part 1.

Context, Intent and Affinity…the “CIA” of Internet Advertising

From a marketer’s perspective, the efficacy of any messaging medium boils down to its ability to:

• Reach the widest possible consuming population. In Internet advertising, the most common metric for reach is the number of unique visitors delivered by a site or an ad network per month, expressed as a percent of all unique visitors on the Internet for that month. For example, Facebook’s reach in the US is about 21% based on the 42M visitors that use the site divided by the approximately 200M total Internet visitors per month in the US.
• Target actual consumers/buyers in the population. This is often very tricky or impossible for traditional broadcast media like TV, Radio, Newspapers etc to do, and is the most cited reason for not being able to measure the ROI for advertising on traditional media. Internet advertising was supposed to solve both the targeting and the ROI problem. ROI is definitely getting to a point where it is quantifiable for Internet advertising, but targeting, while better than traditional media, has a long way to go.

Estimates of eBay Conversion Rates

Recently, I saw an investor relations presentation by eBay (available on the company website), where it provided a nice formula for how the eBay Marketplace makes money.

Briefly, eBay’s Marketplace revenues are a percentage of the Gross Merchandise Volume (GMV) which is the value of the goods actually sold on the marketplace. The relevant equations:

(1) eBay Marketplace Revenues = GMV ($) * take-rate (% cut for eBay)
(2) GMV ($) = Listings (#) * Conversion Rate (%) * ASP ($, average selling price)

eBay publishes GMV and Listings data, but does not disclose conversion rates or ASPs for any of its categories. Listings volume went up by 3% in Q2′08 over Q1′08 after eBay allowed fixed price listings at a drastically reduced 35 cent per listing insertion fee. GMV, dropped in Q2′08 by about 2%, which implies that the conversion rate and/or ASP dropped in Q2′08. Since eBay does not disclose either number, I wanted to see if I could estimate them based on reasonable assumptions.

The US Internet Advertising Landscape

The US Internet Advertising Landscape: Part 1- Overview and Top Level Opportunities

In my quest to find a single report that would teach me all about Internet advertising, I ended up creating one myself by piecing together disparate pieces of data from many sources. As the saying goes, the journey was its own reward in this case.

I will share what I learned in the past few days, in a series of articles, starting with this one.

2008 US Ad Spending by Select Media and Formats

US Internet Advertising Market Segmentation

One of the things I realized very quickly is that lay definitions of Internet advertising include only marketing messages that are delivered on web pages, the usual suspects being text ads for search results pages and banner/text ads on content pages. I suspect marketers don’t really think that way anymore. Rather, a marketer’s problem is really about achieving awareness and/or conversions in the most cost-efficient manner using all available channels that have mindshare with consumers, including:

A Moppet Medium

For a while, my 4-year old and I have been playing a game where we take turns being Barack Obama and John McCain. After Mr. Obama won, my 4-year old wanted to be Barack and Barack only. Not to be outdone, I insisted that she had to make a victory speech since she had won. She asked me what I wanted to hear. I told her that she had to address the country and tell us what she would do for us now that she was going to be President. She reflected for a moment, then made a sweeping gesture and said “Magic”.

On another occasion, I told told her that she could touch my laptop only if she was a product manager. I asked her if she even knew what product managers do. She responded with “they run around, dance around and make everyone happy”.

Chasing the Tail: A Web Search Playbook

A while back, I had argued that tail queries were an important driver of success for search engines (see “Google or Yahoo: Tail Quality Matters” ). In this article, I hypothesize about the characteristics of consumers who drive these tail queries, and how to leverage them for search share gains.

In “Should You Invest in the Long Tail?“  (Harvard Business Review June 2008), Anita Elberse  provides the following consumer insights for online movie rentals:

• Consumers who rent off-beat (ie tail) movie titles tend to be heavy consumers of movies overall.. As the red line in Figure 1 shows, consumers that rent the least popular movies (movies in lowest deciles) consume about 50 movies per month versus 20 movies per month for the top deciles.
• Light consumers tend to stick with popular titles
• Off-beat movies tend to be rated lower than popular titles as the blue line in Figure 1 shows. In other words, consumers are less satisfied with off-beat titles than popular titles, despite seeking them out.

While I have not seen similar data for web searches, should these insights carry over to web searches, how would they change strategy for search engines, especially the ones that compete with Google? Before going further, let me re-frame Elberse’s insights as hypotheses regarding web searchers:

• Heavy searchers (those that do more than N searches per month) dominate tail queries
• Light searchers (those that do less than N searches per month) stick to head queries. Anecdotally, shopping, travel, local (e.g. restaurants), do-it-yourself (e.g. recipes, home improvement, gadget repair) and entertainment (movies, music etc) related queries account for the bulk of head queries.
• Searchers in general are less satisfied with tail query results

Search Strategy Playbook

• Winning tail query shares requires search engines to focus on the needs of heavy searchers. This follows from the hypothesis that tail queries arise largely from heavy consumers. Possible strategies include:
  • Open up search results so that heavy searchers become contributors/enhancers of search results. Given the hypothesis that consumers are not happy with tail query results, and the fact that sites like Wikipedia leverage consumers to contribute content in a self-policing way, this strategy might actually work. In my own experience with tail queries, I can usually get to what I am looking for with some amount of digging around in Google and Yahoo!. There should be a way for a search engine to leverage such work by letting a searcher either rank the most relevant result or submit a result set from a different search engine. Subsequent requests for these queries should factor contributed results and/or ranking in addition to the usual relevancy criteria.
  • Personalize the discovery of query intent to deliver more relevant tail results. There should be a way where a heavy user can specify domains of interest for tail queries so that search results can be tuned. Start-ups like Rollyo are beginning to do this.

• Non-US markets appear to be an open playing field for all search engines- again, opening up search results will be key to success. In these markets, head queries themselves (shopping, travel, local (e.g. restaurants), do-it-yourself (e.g. recipes, home improvement, gadget repair) and entertainment (movies, music etc) are the initial opportunity. For example, finding information on a destination like Rome required me to go to multiple travel sites and blogs, after I was unsuccessful with popular search engines. A strategy that lets searchers (heavy and light) to rank, contribute or augment search results would go a long way to boost search market shares for engines that allow it.
  • Crawling links in social bookmarking applications such as Del.Icio.Us within a searcher’s social network should also be useful for Non-US head queries. This will work particularly well for head queries such as recipes where social bookmarks would be quite valuable.