Matthew Effect: The Motivations of Google+

Posted on 23. Nov, 2011 by in Thoughts



How the “Matthew Effect” and the move to attribute influence to authors i.e. people, is the motivation behind Google+, and why Facebook is only a commercial rival to Google+.

Introduction

Since its conception Google+ has been compared to Facebook and thus seen as a rival in social networking. I don’t believe Google+ is a rival to Facebook. I agree Google+ and Facebook share similar characteristics such as networking, liking, profiles, etc, but I believe Google+ is a search enhancement that better enables Google to capture discrete interactions (engagement) that culls the “Matthew Effect” economy that persistent interactions (citations) have created – a legacy issue also noted when measuring scientists and scholars.

Matthew Effect

The Matthew effect is the phenomenon where “the rich get richer and the poor get poorer”. Those who possess power and economic or social capital can leverage those resources to gain more power.

“For to all those who have, more will be given, and they will have an abundance; but from those who have nothing, even what they have will be taken away.
-Matthew 25:29, New Revised Standard Version.”

I believe those who hold the former, power and economic capital, currently rule Google’s search results – natural and paid. This makes Google unfavourable for the “poor”. Google won’t allow search to exist in that manner; it goes against the grain of Google – empowering those based on merit and to be seen as an open marketplace.

In the opposite corner, social capital is not ruled by those who hold power and economic capital. Social capital gain is the basis for Google+, creating a more open playing ground for the “rich” and “poor” to compete again. Partly at least; it opens Google’s market-depth by changing parameters for which websites rank, which entails that Google will become more appealing to mid-tier and SME (Small and medium enterprises) by leveraging social capital in the form of people as a ranking factor.

Asymmetric vs. Symmetric Social Networks

There are different types of social networks, asymmetric and symmetric. Google+ defines itself as an asymmetric social network.

  • Asymmetric – An asymmetric social network is a social network in which a first member’s relationship to a second member is not necessarily the same as the second member’s relationship to the first member. Since the character of the social interaction between members in a member network can be defined in accordance with the nature of the relationship between those members, a first member in an asymmetric social network may interact with a second member in ways that differ from the social interaction provided for the second member to interact with the first member.
  • Symmetric – A symmetric social network, related members necessarily share the same relationship with one another. Examples of such symmetric social networks include Facebook, LinkedIn, and MySpace, where two or more members establish bidirectionally equivalent “friend” or other relationships generally using an invitation/response protocol that effectively requires the consent of both members to the relationship. Such bidirectionally equivalent relationships provide the same social interaction possibilities to the related members.

Use this table of asymmetric and symmetric social networks to help understand the difference:

Asymmetric Symmetric
Google+ Facebook
Twitter LinkedIn
  MySpace
  Bebo

The main different between both asymmetric and symmetric is the bidirectionally equivalent. That being, with asymmetric social networks “a first member may be a follower of a second member without the second member necessarily being a follower of the first” – where symmetrical social networks, you have a commonality (bidirectional equivalency) e.g. you’re friends, co-workers, family, at the same gym, etc.

Google+ defines itself as an asymmetric social network. Looking at its features, Google+ is an asymmetric social network at heart, but with symmetric features. You can follow a user or business without them necessarily knowing who you are (bidirectional equivalency), while at the same time Google+ uses circles to classify your connections.

Following a user and having them following you back (reciprocal) makes Google+ asymmetrical, but put the user in a circle called “friends” or “co-workers” and Google has its bidirectional equivalency that symmetrical social networks hold. That is a major difference between Google+ and Facebook. That said, Facebook allows people to “subscribe” to their feeds, which is asymmetric.

Who it favours?

The important point to uncover here is that being both asymmetric and symmetric makes Google+ favour businesses, which seem to prevail on asymmetric, and people who prefer symmetric.

Out of the two, Google+ sways heavily towards asymmetrical. That is a signal that Google wants to empower the “little guy” we’ve talked about. People can prosper without having to have a symmetrical bond.

Persistent vs. Discrete Interactions

Google has only since been able to level-peg the search playing fields of the “Matthew Effect” with the introduction of Google+. Google+ allows Google to capture signals beyond persistent interactions, leveraging discrete interactions for competition between both the “rich” and “poor”.

According to research from IBM entitled “Towards simultaneously exploiting structure and outcomes in interaction networks for node ranking” persistent and discrete interactions can be classified as:

  • Persistent – weblinks, friendship, membership, affiliations.
  • Discrete – e-mails, collaboration, authorship, team work.

Google+ opens up discrete interactions to Google. Arguably, discrete interactions (or engagements) now fuel 35% of Google search queries i.e. fresh.

Google’s PostRank was the window

Google’s acquisition of PostRank was the window of opportunity to capture discrete interactions. PostRank defines discrete interactions as “engagement signals”.

On PostRanks website, there is a list entitled “The 5 Cs of Engagement”. These include:

Creating -
The strongest form of engagement is demonstrated by using an item as inspiration to create your own, for example, writing your own blog post that responds to or refutes someone else’s blog post. Creation requires the most thought and investment of time, actively generates conversation, and therefore indicates the highest level of engagement.
Critiquing -
Reading a blog post and then leaving a comment requires an investment of time, thought and effort (or sometimes just typing and name-calling…), and is a form of conversation. However, it requires less effort than writing a whole blog post. So while it is an important action, it does not indicate as much engagement as Creating.
Chatting -
Sharing and discussing information can often be started with one click, so it doesn’t require a major investment of effort. However, a desire to share is a strong indication of relevance, and the act of sharing and its ensuing discussion are acts of conversation. Use of social media applications like Twitter encourage both the sharing of information and the resulting conversations. As a result, social media “chatting” indicates a good level of engagement.
Collecting -
Bookmarking or submitting items to social sites also tend to be “one-click” actions. They are intentional acts of archiving and sharing, but don’t require much time or effort. However, the sharing that occurs often sparks conversations, so Collecting does demonstrate some engagement.
Clicking -
Activities like clicks and page views indicate lower engagement because they’re passive interactions. Clicking a link to read a blog post doesn’t require much work, and you’re not giving anything back except your reading time. It is an intentional act, however, and thus indicates a mild level of interest and engagement. Which may grow after the item is read.

All of PostRanks “5 Cs of Engagement” are interactions attainable from Google and Google+. If you chat, critique, collect, click, bookmark, share, comment, create, write, collaborate – if Google can – it will – capture those signals and factor them into the search results.

People have argued for a long time whether Google uses click-through rates (CTR) and Google Toolbar data to help rank the search results e.g. “clicking”. Owning PostRank makes that possible. Personalised search gets more exciting.

What about Bing?

Microsoft’s Bing recently went public suggesting they rank webpages based on:

  1. Content
  2. Social
  3. Links

Dissolving Microsoft’s list into PostRanks “5 Cs of Engagement” and we can compare Bing and Google’s engagement ranking signals:

  • Google – Creating, Critiquing, Chatting & Collecting (Social), Clicking (Content) = 2
  • Bing – Content (Clicking), Social (Chatting & Collecting), Links (Persistent) = 4

Note: Bing captures 2/5 of the “5 Cs of Engagement”.

I think Google has invested more in discrete interactions with Google+ where Bing would appear on the surface more limited in the engagement it captures – by not having a social network (i.e. Google+). Bing does use Facebook data, but it comes at a cost and is limited – likely only “Chatting”, “Clicking” and “Collecting”. It owns Skype, which would host a vast array of signals – not linked to authors, but nevertheless, signals. Skype is an ace card for Microsoft.

Google Was Only 3/4 Complete

Until Google+, the search engine was only ever 3/4 complete (or baked). Google was able to use PageRank to rank documents based on how authoritative they were deemed by looking at citations i.e. links, amongst other persistent signals, but those efforts have never allowed Google to reach 100% – such signals could be controlled by power and economic capital. For example, paid links, affiliations, money, etc, favoured those with power and economic capital. This deters the “poor” from Google – which isn’t an option, because it limits Google’s potential reach commercially.

35% of Queries are “Fresh”

Recently, Google cemented its commitment to social capital and the “Matthew Effect” by promoting fresh, new and recently edited content to the forefront of the search results.

Using persistent signals such as links does not permit Google to achieve this alone. For example, Joe Blogs with a small time blog could uncover a hot story and within minutes of publication, a major newspaper or established website with large power and economic capital could cover the same story, and due to their persistent signals, Google’s algorithm would favour the website with the most power and economic capital. That has since, from what I can see, started to change, leveraging social capital and what I would describe as “the little” people to compete for arguably, at least 35% of queries that are “fresh” – over 1/3rd of the pie.

Ranking Scientist & Scholar Effort

At the heart of Google, science and scholars have always played a role. PageRank is based on how to rank documents based on citations; historically scientists and scholars have been ranked in the same manner. Google Scholar is testament to their commitment.

Beyond ‘just’ citations, there are various other ways to rank individuals, like how frequently they publish a paper vs. first paper published or more interestingly:

  • how to score individuals on collaborated work
  • the impact of their work
  • recent authority within a field
  • how productive one is or
  • how a group associated with a organisation perform

All of list described are features of the following algorithms:

Some are based on PageRank e.g. SCImago Journal Rank, others pre-date Google itself. That aside, such ranking algorithms have been used for years. However, they all fall short of perfection e.g. 3/4 perfect, just like Google.

The final hurdle in my eyes was for Google to decide how to attribute discrete interactions (engagement) to authors – not “websites”, and the role authors play in content production. I think Google+ enables that – which none of the above have been capable of doing.

Take h-index for example, the most similar to Google+ in terms of measuring authors:

The h-index is an index that attempts to measure both the productivity and impact of the published work of a scientist or scholar. The index is based on the set of the scientist’s most cited papers and the number of citations that they have received in other publications. The index can also be applied to the productivity and impact of a group of scientists, such as a department or university or country.

Keywords from the description being “measure”, “productivity”, “impact”, “most cited”, “number of citations”, “other publications”, “group of” and “department”. Translate these terms and apply them to authors and how Google should handle author authority, you can see why Google+ and discrete signals are important. However h-index has criticisms:

  • Doesn’t account for the number of authors per paper. Doesn’t partition citations amongst authors.
  • Doesn’t account for citations per different fields. Different fields tend to leverage more or less citations. H-index doesn’t take into account.
  • Doesn’t consider placement in author lists, which is significant in most fields for how much was contributed per author in a top-down manner.
  • Bound by the number of publications. Those with short careers, regardless of importance of the publication, are disadvantaged.
  • Context of citations – used flesh out the intro, are negative, or false. H-index ranks all equal.
  • Books vs. articles; distinction between different types of publication.
  • Gratuitous authorship is a problem e.g. Matthew Effect.
  • Natural vs. rational number used to compute scores. The latter is preferred.
  • Self-citations e.g. spammers using the likes of SCIgen to create nonsensical citations.
  • Doesn’t normalise scores per expertise.

I believe such criticisms of h-index are solved in Google+ by how it captures discrete and author interactions. For example, Google+ can:

  • Author tag allows one to define authors
  • Google can classify authors per field, or more explicitly (or important), by search keyword.
  • Google has long valued links in an order of important since using Yahoo! Directory.
  • Career length doesn’t matter as both fresh (35%) and stale/static (65%) content can perform.
  • Sentiment can be estimated. Negatives citations are often good results for Google. Im confident Google can also rank citations based on where they appear, and what content.
  • Google+ can and does distinct between books and articles with Universal search e.g. blog, news, web, etc.
  • Self-citations are easily de-valued by Google’s investment in Search Quality.

As the final quadrant of the pie, Google+ and the signals it captures from Google+ and the search results (and beyond) makes Google’s search more complete (for now).

The Nobel Prize

When Google entered the search engine business, it has always been wise in combating not only the “Matthew Effect”, but being able to attribute merit where merit is due.

There are three cases where the Nobel Prize has been awarded to scholars whom may not have should rightfully won. These include:

  • George Sundarshan – another scholar won the Nobel Prize, when George Sundarshan achieved the same work on the same topic but was ignored.
  • Nobel Prize in Chemistry – award was given to a prominent relative newcomer and previous work on same subject ignored.
  • Memorial Prize – was awarded to one scholar where another published identical work the same year. It should have been shared.

If the Nobel Prize were awarded based on who topped Google’s search results, it could and would have combated all of these cases above. That isn’t a fair statement because Google has 10 winners, but using a combination of “fresh” content and discrete interactions (engagements), Google would be able to determine which version be ranked first, by not just looking at author authority.

For example, if X published their research first, Google may favour to rank that content above more recent research of the same topic. Additionally, if two scholars produced the same content, Google would value the research that other scholars engaged with. Sceptics whom complained to the Nobel Prize about the wrong scholars being awarded could be proven right or wrong using Google’s search.

I highlight the example merely to show the historical issues of trying to reward authors correctly.

SCIgen

Scholars are spammers too. Using randomly generated content, SCIgen can be used to create nonsense research papers with accredited authors. Create enough nonsense papers and your authority in your field would increase.

This occurs during today’s web – mainly by those with large economic capital. Looking at beyond the persistent interactions and paying more attention to people and how they discretely interact e.g. share, converse, engage, etc, makes it easier to detect and discount SCIgen on Google.

Expected Ranking based on Structure and Outcomes of Interactions

In the research from (PDF) IBM referenced before, their goal was to devise a solution that ranked content based on both structural property and outcomes of interactions – I translated that into digital marketing speak as ‘trying to provide a solution that ranks content based on both connections and engagement signals’.

“…our goal is to develop a ranking scheme that takes structural properties and outcomes of interactions. Consider the interaction network of Figure 2 arising out of interaction from figure 1 with outcomes II (revenue generated) from the domain of service interactions. Let the weight on each edge be unity. We will use this example to highlight some of the deficiencies of past approaches. We will begin with an ideal ranking that takes both structure and outcomes into account.”

Within the research, IBM uses a dataset of IMDB actors. Using Meta data associated with each actor such as the number of movies starred in, fellow cast, movie revenue, etc, they re-enacted a social network – asymmetric and symmetrical – looking to define author authority. Simply replace actors and Meta data features with that of a social network e.g. followers, friends, likes, comments, tweets, etc and you should get a better picture of the study – and hopefully value its findings.

Without going into too much detail, IBM found that by altering the data associated with actors such as the movies they’ve co-starred in impacts the score or author authority of the action in the symmetrical relationship. Here are two examples I thought up:

  • Upcoming Actor – new, upcoming actor has a relevantly low author authority. Have them star or play a role in a movie with Robert De Niro, and the influence Robert De Niro has on the upcoming actor increases their authority. Applied to Google+, upcoming person writes a piece which is published on Mashable, that person’s author score will rocket. More granular, the upcoming author contributes to a piece on Mashable with two other established authors – score goes up.
  • Modifying Data – by changing the gross revenue of “Forrest Gump” from $625 million to $1,000, Tom Hanks score would change impacting his author authority based on how heavily weighted gross revenue was.

What I loved about this research is that it’s exactly how I envisage Google capturing and measuring authors and people on Google+. Replace actor attributes with engagement and social signals and you can see how it applies.

IBM’s research suggests it only works on a symmetrical basis. I believe circles make it applicable to Google and its asymmetrical roots.

Conclusion

I concede that Google is a social network. I also concede that Google+ is a rival to Facebook, primarily due to the scale and nature of both commercial beasts.

I do however hope I have put together a useful enough case to suggest that Google+ isn’t bidirectionally ‘just’ another social network, but rather a very powerful enhancement to its fundamental business – search and advertising. I hope that becomes more apparent over time, because my initial impression is that Googl+ has been ring-fenced as social media and not a search engine marketing tool. That would be a mistake for any business or personnel to make.

I would also like to think that the “Matthew Effect” described is dissolved with people power, and people like you and I, not faceless brands, better establish themselves as valuable assets – Mashable’s CEO Pete Cashmore sums up how authoritative people can become. Expect to see many more mini-Pete’s prosper along with Google+.

People power and the modifications in ethos is a wise move from Google, not only because it solves and improves search (rule number 1 of fight club) but because it makes Google a much stronger marketplace for all forms, shapes and sizes of advertisers, filtering into Google AdWords, Product, Analytics, etc.

I the words of Google themselves (video):

“Business don’t make people happy. People do”

A motto I’ll be amplifying in any social search strategy Im involved in.

What do you think, is Google+ a search and commercial enhancement or means to compete with Facebook in social networking? Feed the engagement, share your views via Google+ or Twitter.

Tags: , , , ,