Facebook and Social Search: The Database of Suggestions vs the Database of Intentions

I was at a social networking networking event (I know... :-)) recently and ran into a couple of guys from one of the public UK broadcasters. When asked why they were attending, they mentioned their website had been getting a lot of hits from Facebook and they were interested in finding out how they could encourage this. This was real world confirmation of the trend which began with Digg. Social recommendations are becoming mainstream, and organizations are trying to capitalize on it.

Social recommendations are tricky and all involve an element of trying to get your brand or product to go viral. Advertisers have to invest upfront in apps or games (such as the Whopper Sacrifice) or create compelling profile pages and then hope that they take off by word of mouth. It is much harder to pull off and target than simply buying AdWords on Google. However, a combination of search and social recommendations, that advertisers can measure in a CPCs has a chance of giving Google a run for its money. This is the premise behind social search.

To date, Google's dominance in search has not prevented new search engines from springing up. New entrants have differentiated themselves by sector (e.g. job search or housing), region (Yandex, Baidu), presentation (Kosmix) or ability to answer specific types of questions (WolframAlpha) amongst others. Those that have tried to compete head on (such as Cuill) have not fared too well. All had one thing in common - they have access to the same data as Google, which makes it that much more difficult for them to compete. How does social search differ?

Most importantly, social search has the potential to be highly targeted. Facebook, MySpace and others collect data such as their users’ comments, likes and dislikes and limit (or allows users to limit) what is publicly accessible. This additional data, should in theory makes search results more accurate – even Google agrees with this. Due to this, the blogosphere is awash with articles guessing how social search is going to affect Google's dominance. A recent article in GigaOM compared Google's search to looking for things in the library, and Facebook/Twitter driven search and recommendations as exchanging ideas in a coffee shop. While Google is great at indexing all published articles and can use your past results (if you let it) to make intelligent guesses to your current needs, it is never going to be as good as your friends' recommendations. To put it another way, if Google is the Database of Intentions (to quote John Battelle), then Facebook/Twitter could very likely form the Database of Suggestions.

The flip side of the coin is that determining the relevance of social recommendations may be quite difficult. Visitors to Google are signaling their intentions by making the effort to search for something specific. Users of Facebook may be driven by curiosity after finding a link on their friends' wall. Unless it is possible to filter this traffic, advertisers will have difficulty getting consistent ROI. Presumably, Google with all its PhDs should be able to come up with a model that works? Google has had a fair crack at this through its partnership with Fox, powering MySpace search amongst other things. However, Google executives have from time to time let it be known that this deal has not met expectations. A recent TechCrunch article provides actual numbers.

Besides relevance, there are also concerns about personal privacy and the damage to brands as dynamic user generated content (UGC) impacts the ability of advertisers to filter ads tied to objectionable content. However, the benefit of doubt has to be with the social networks. With an audience of 100s of millions and ranking in the top 5 sites in most countries, they clearly have share of eyeballs. Advertising dollars have to follow and this will stimulate innovation. They are also the victims of high expectations - after all, how many years did it take for Search 1.0 to evolve to produce an AltaVista or Google?

While social search is still on the horizon, social recommendations are already having an impact on SEO practioners and agencies. Agencies are beginning to talk about multichannel internet marketing, increasingly using not just SEO and SEM but also blogs, social networks, Twitter and others as a source of traffic. It will be interesting to contrast these different methods against SEO and to explore the effect they will have on the digital agency business model.

Game Theory and Software Business Partnerships

A recent discussion on game theory and sequential games triggered this article.

Managing partnerships with multiple players in an organization is a challenge, regardless of the economic situation. However, a down economy imposes specific strains on partnerships as organizations become even more focused on short term objectives. Projects that were in the process of getting into a growth trajectory are in danger of being pared back or terminated as the organization goes into survival mode. In addition, companies are subjected to disruptive events such as layoffs, which rob an alliance of the relationships that have been built over long periods. 

What can a partner manager do to ensure that their organization does not make an investment that is suddenly rendered useless as a result of a cancelled agreement? Game theory has something to teach us here.

As business partnerships are legally constituted and involve multiple interactions, their equilibrium state can be modelled as cooperative sequential games. However, if either or both partners become aware that the relationship is about to end, i.e. that the next game could be the last, they revert to looking after their own needs, resulting in a suboptimal outcome for one or both parties. What can a manager do to prevent this situation occurring? The answer is to engineer a contract that encourages ongoing cooperation and penalizes (premature) termination.

Some options:

1.     Deferred payment: In a deferred payment situation, one of the partners is a manufacturer or vendor, and the other a reseller. The reseller collects a fee immediately and is paid a deferred incentive. The incentive can be paid over multiple years and made contingent on a minimum number of sales per year. It can also be increased for each period where the vendor outperforms. If the relationship terminates prematurely, the deferred amount can be withheld. The net effect is that a successful reseller that brings in multiple deals, will have an increasing incentive to continue the relationship.

2.     Options: Options can be given over a period of years and programmed to vest only if the partner is meeting a baseline. Options can be more powerful than revenue as they give the reseller a stake in the overall business success of the vendor, without the risk of a merger. The risk of vendor underperformance remains however. A famous case is that of the early advertising deal between Google and AOL, which worked out extremely well for AOL, and was also the deal that ‘made’ Google.

3.     Sharing board members: A famous example is that of HP and Cisco sharing board members. Carly Fiorina, then CEO of HP, was a Cisco board and during her tenure, HP emphasized Cisco’s products over it own. Weaker commitments can be made using shared non-executive directors.

4.     Vendor Lead Program: A strong lead generation program by the vendor can ensure that the relationship is two-way. Apportioning high quality leads to resellers that deliver the most sales is a great quid pro quo and can be a powerful motivation for them to stay in the relationship. This is especially useful if the reseller has multiple offerings as each new customer has a value over and above that of the immediate (vendor) deal. In these types of programs, it is important to track lead conversions and review the program with executive sponsors.

5.     Joint marketing program: The vendor and the reseller agree to run marketing awareness activities on an ongoing basis. This could include marketing events such as customer workshops, with incentives for meeting targets on customer attendance. This is a variant on the vendor lead program.

6.     Shared employees/infrastructure: Vendors can offer to sponsor employees dedicated to their products at the reseller premises. Employees are jointly recruited and trained and their remuneration is shared. Having made an investment, neither party is likely to want to prematurely terminate the relationship. This can be repeated with infrastructure, for example a testing facility, which is specific to the vendor.

The litmus test for any partnership is value creation. The above measures are not substitutes for this. However, in a partner situation with multiple priorities and limited resources, a calendar of ongoing revenue and defined activities will win out over short-term expediency.