When you start talking to an enterprise crowd about social networking, inevitably someone asks for real business benefits, a.k.a. ROI. I know, hard to believe.
When we first started the ‘Lab, Paul used to ask how many people have a Facebook/MySpace/LinkedIn account. Usually less than half the room would raise hands, and that number went way down when he’d ask how many use and update those accounts regularly.
Now, Facebooking is a pop-culture punchline; everyone does it, keeping the status updates as current as possible, checking for news feed nuggets, making new “friends”, throwing sheep, all that. So, as more people use Facebook (recently reaching 175 million users), more people get what social networks are about and get their appeal.
Real value to everyday work remains hazy though because most companies don’t have internal social networks, and many actively discourage, if not block, the use of consumer social networking sites during work hours.
Even businesses that allow employees to use social sites see them as nice-to-haves, easy ways to keep employees satisfied with work. Sure, employee satisfaction is a business benefit, but as the layoff parade continues, employee satisfaction can quickly be boiled down to having a job at all.
One must-have facet of work, regardless of the type, is the ability to find information.
Information, and the need for it, applies across the board, in all manner of work, blue, white or other-colored collar, and the more information we have, the harder it becomes to find the right information. So, we need good search, which is defined as good results.
Let’s use the ‘tubes as an example, without search, how much would you use it?
Think about how frustrating it would be to use the ‘tubes if there weren’t a Google (or other search engine) to help you.
The inability to find information within the firewall must be one of the top five highest ranked problems according to CIO types. People constantly complain that they cannot find this or that, and this is the type of problem that keeps CIOs up at night because there’s no easy answer.
Enterprise search has loads of players, but each one is constrained by iron. In fact, many times people assume that enterprise search is no good because it never has a chance to shine. Flipping back to the ‘tubes, has anyone else noticed the enormous investments in big iron and data centers that Google has made over the last few years?
Someone is. Here’s a map. There are nearly 40 Google data centers in the world.
This represents hundreds of millions of dollars (probably billions) invested in computing power. Not the kind of investment an enterprise can justify to help its employees find that pesky pdf with the benefits enrollment instructions.
Google, as the standard for search, has set the bar for response time very high, even if its results aren’t always the best. PageRank is a good algorithm, and Google is constantly tweaking it. However, the emphasis on speed has been the key to Google’s continued success.
Enter search as the benefit to social networks, i.e. social search.
Social search is an area where enterprise can pwn consumer web. Social search, i.e. filtering search results based on your friends searches, comments and rankings, would be downright creepy on the ‘tubes.
However, at work, it can be very effective. Let’s consider the path of least resistance for questions.
In my experience, questions go to people first, over the cubicle wall, over the phone, over email, over IM, and lately over Twitter (or OraTweet). Always to a person though, not to an algorithm.
Why? Because we know each other in the context of work, e.g. I’m known as a member of AppsLab, therefore I get questions about Mix and Connect and blogging and Web 2.0, etc. I’m tagged people’s minds, even before they use a social network. Before joining the ‘Lab, I was in EBS Financials. So, I had those tags.
It’s human nature to associate information with people.
Social search is based on people’s knowledge, but rather than pay editors (like Mahalo), why not use a social network to collect people’s rankings of results? So, when I search for “benefits” and find that pesky web site, I can comment on the result or “like” it or bookmark it or rate it or in some way indicate that it’s important to me.
This action is broadcast to my network and always visible to them. That may not matter until Rich is searching for that pesky site. His search for “benefits” when filtered or weighted by a social dimension will bubble up my results. This should save him the trouble of paging through results, which is a tedious and short-lived process.
Most people don’t even make it to the bottom of the first results page. I don’t. If I can’t find what I need above the fold, I either ask someone (IM, IRL, email, Twitter) or modify my keywords to try again.
I don’t think I’m unique here.
Twitter is quickly becoming a great way to ask questions of a large audience. As more people use Twitter for questions and advice, and because it’s all indexed, Twitter search is becoming a great way to get answers without even asking a question.
This message plays well to enterprises for a couple reasons.
- Social search doesn’t need big iron like enterprise search does.
- You could make a good argument that social search is as good or better than any algorithm because of human cues that cannot be modeled.
- Social search is useful inside the firewall because of inherent trust and commonality of purpose. It’s creepy on the ‘tubes.
- Social search is safe inside the firewall, or at least safer.
This is a real business benefit of social networking, and there’s plenty of ROI to be found in fast and accurate access to information.
What do you think? Find the comments.
that just about sums up my feelings… technology cannot replace people, enterprise search is hard, and the “semantic web” is a joke compared to tools that allow “social search.”
The enterprise may never have the data centers that Google has… but it can use identity management and data mining tricks that Google NEVER could get away with to help jump-start social search. The enterprise also doesn't have to worry about SPAM…
Of course… all is for naught if you don't also institute gatekeepers, or some kind of valve to prevent “social overload.”
I agree that social can help improve search, it adds trust to data. One big problem in the enterprise is that data is hidden behind closed applications. This means that a search engine need plugins/connectors to get to the data. This is quite complex and expensive if you have a lot of different apps.
The internet is/was different in that all the data is basically out there for you to grab it, embedded in html. Unfortunattely, also on the internet we're more and more hiding data behind closed applications (ajax, flex, silverlight).
To really be able to index all data, we need an internet compatible standardized data access layer. Web Services aren't the solution, as each service has it's own interface. Another problem with Web Services is that the data isn't really crawlable, all data more or less exists in it's own little world.
I think REST can solve this problem: internet wide id's, links between data to be crawled, standardized API to query data.
* http://www.andrejkoelewijn.com/wp/2008/12/15/wa… (Want to find your data? Use REST)
* http://www.andrejkoelewijn.com/wp/2009/02/12/re… (Rest is a distributed data model)
A practical example from 2 weeks ago.
Someone asked me (via email) if I had a contact name for the developers of Aria (Oracle's People Search Tool). After toiling through our intranet and not having any luck. I posted the question on OraTweet, and had an answer within a few minutes.
A colleague had the same experience when trying to find information on Newsletter best practice. He sent me an email, and as I didn't have the answer, I directed him to Connect. By the end of the day he had a few leads to chase up.
So from my perspective Social Search is definitely becoming more popular.
Yeah, reading over my post, I'm planning a part 2 today. I didn't get into the meat of it very much. You touched on a couple points I didn't emphasize enough, no spam and trust.
Agreed. Structured data are much tougher to index. I'm not suggesting a full replacement of search engines either. They're still needed to get initial results that are then refined by social metadata.
I like your suggestion about using REST. I plan to do another part on this today, and I'll check them out before I put pen to paper 🙂
I'd argue it's always been the most popular way to get answers, but physical limitations have forced people to use other methods with some success. New tools have opened up the weak ties we already have, allowing us to mine them.
Your use case is very similar to one Paul had with Connect. These are solid gold for people who need use cases showing ROI. We should collect them . . .
I think this social networking is great a way to meet new people
and get up to the minute information and it's free all it
cost your time