Does technology need to solve everything?

I needed soup. I was coughing, sneezing and dying a slow death called body ache. I needed something warm, something soothing. A few moments of relief. Soup.

It was my first year away from home. I had never bought packaged soup before. What I bought was trash. Desperation. All I could do now was Whatsapp a friend. ‘Dude, any restaurant in the vicinity that serves good soup?’

He said ‘Why would you want to order soup from a restaurant. Go to the Supermarket and get soup brand X. Trust Me’.

Good soup procured and the rest was history.

So yesterday when I read about a start-up providing social recommendations for shopping, I totally saw myself as a user. Recommendations from your close friends sometimes make all the difference. Even for something as mundane as soup.

The product was simple. Users keep recommending products that they feel are worth recommending. Using your Facebook data, the system figured out whose opinion mattered to you so that when you needed recommendations, you’d have them at your fingertips.

Great! Sounds like a plan.

Genuine need. All the right pieces of Technology — Mobile apps on all platforms for easy access, Social Mining to find people who matter and finally a platform which is time-persistent to record recommendations and provide them on-demand. Technology solves everything. Yay!

So now I need to buy a new-T-shirt. Friends, what be your recommendations?

Wait. Why would any of my friends keep adding recommendations to this service?

In real-life friends help-out each other because ‘They Care’. When you give them a call or text them, they do their best because they genuinely want to help you out. How do you translate ‘Genuine Care’ to a functionality?

How do you make caring an app-based incentive? Why would a user spend time recommending products on a third party service without any immediate needs or gains? We can always give them coupons but then we have enough of those anyways.

Secondly, when you need recommendations, who would you get it from?

Probably this friend who knows about where all the offers are. Or maybe that friend who has those ‘Superhero T-shirts’. Or maybe that friend who has the same shopping budget as you.

How do you translate these to code? How do you record a user noticing a superhero T-shirt that some friend wore a couple of weeks back and use it to provide recommendations?

Sure Facebook has a lot of data about your friends and acquaintances. Pinterest knows the people whose collections you like. But the way we seek recommendations is more impulsive and intuitive than what data can currently model today.

And lastly, how do we actually seek these recommendations?

Do we like just ask a friend for some store names and then hang-up?

Nope. The process is much more than that. It starts with your friend asking you why you need something, what your criteria is and maybe how much you are willing to spend. After you've answered these, your friend makes tailored recommendations for your needs and then goes on to add some extra goodness like which brands to watch out for, which tailor to insist for and some real good advice to drop a criteria or to increase/decrease the budget.

Social recommendations the way we seek them today are much more conversational in nature than just reading a page full of information. It’s better to give your friend a call than going through 30 pages of information in the name of recommendations. It’s a no brainer that a call or text messaging is the way to go.

When I think of such questions, I realize that our ways of doing some things are much more intuitive than technology can offer.

Maybe some problems should be left alone.

Sometimes, the best way is the old school way.

LinkedIn - Making Deeper Connections

'ARE ALL MY LINKEDIN CONNECTIONS THE SAME?'

Last week I said hello to LinkedIn after a really long time. Trust me when I was in my final year, this used to be a lot more. While accepting the numerous requests to connect, I took some time to actually go through my LinkedIn connections after a very long time. Normally, I get LinkedIn requests from a broad spectrum - people who I genuinely know, people who I haven’t interacted with much, people who I know just by name and even complete strangers, same as any other LinkedIn user. 

My process is simple – check if I know a connection or if I have something in common with a requester or lastly, if there is any value to adding them to my ‘Professional Network’. It keeps my connections clean and includes people who I really want to know about or keep in touch with.

On the other hand, while sending out requests, I usually add people I know well. But mainly, I use it regularly to add folks who I have interacted with, sometimes pretty miserly, at various networking events like competitions, conferences, barcamps and so on, if they are genuinely interesting. It’s like keeping somebody’s business card if you need to reach them someday. Some people I know use Facebook and Twitter for this, but I draw the line at LinkedIn for folks like these.

What I was really pondering was the purpose of adding these connections?

A few obvious ones like below come to mind -

1. Migrate your real network to a virtual network and then extend that network by reaching out to folks in the virtual networks of people in your network


2. Use connections as a reference or an introduction to reach out to and speak to new people of interest


3. Find somebody within a given organization to get referred or probably some information


4. Keep a track of the career changes/milestones/ skills etc. of the people you know so that you are always updated


5. Improve your discoverability to pop up on top when somebody is looking for prospective hires, skill sets and so on


6. Adding weight to your profile by showing off your network capital


7. Get valuable reading content shared by great connections

If you ask me 2,3 and 4 are what people actually trying to achieve. The rest are good to haves.

Now Quick question – I want to get referred to a position at ABC Corp which I think suits my skills. Who would be a better connection to get information about that position and possibly a reference?

• Somebody I met at Robocon 2012 or at IIT Techfest who now works at ABC Corp


• A guy at ABC Corp who my classmate or close friend regularly plays tennis with but I don’t know directly

Unless you have a magnetic or memorable personality or you met a really nice guy at Techfest/Robocon, in the most cases, it will be case no. 2. Which brings me to my original point –

‘Are all my LinkedIn connections the same?’

I feel not. There are ‘buddies’, ‘connections’, ‘acquaintances’, and simply ‘random’.

Though adding all of the above makes absolute sense and provides great networking capital, can we take a step back and think about how to provide a user with the ‘Optimum’ way to perform a ‘Task’?

Starting a conversation, getting referred, getting an expert opinion can be tasks. Though all your connections are a part of your network, not every category of connections will give you the same ‘return on connection’.

A 2^nd Degree contact may a lot of times get a task done much more easily and quickly than a 1^st Degree contact.

Now let’s look a step ahead. How do we characterize, deeper connections from the looser (loser) connections in real life?

You usually have a deeper connection with somebody who you interact with regularly or strategically whereby the person on the other end really gets to know you. This can both be in a formal or informal setting. This can be a guy who you work with on your team, your childhood friend, somebody you play tennis regularly with, somebody who you know at water cooler discussions, etc. Infact, it’s a pretty accepted fact that the closest folks at even offices are the ones who drink or smoke together.

So how do you discover and map these relationships to LinkedIn?

Using something very obvious – People who share deeper connections are also very likely to reach out to each other on other social networks, informal ones like Facebook or Twitter. By periodically accessing a LinkedIn user’s Facebook or Twitter data, one can continuously enhance information about a relationship.

Is being friends on Facebook or Twitter in addition to being LinkedIn connections a good way to term a connection as deep?

I say we go deeper. One can monitor for ‘interactivity’ on these networks. This can be simply comments on each other’s posts, posts on each other’s wall, likes and so on. All these actions specify a different level of interactivity. A comment on a post by the connection can be considered a better interaction than a ‘like’ or a ‘retweet’. Employ a weighted system to derive a quantitative measure of interactivity and assign it to relationships on LinkedIn.

Now an example -  Let us say our system assigns a ‘deep weight’ of ‘1’ to an acquaintance and increases this value on a fixed scale as it finds more interactions on these social networks. Let us say Ram is Shashi’s classmate (Deep Weight = 3) and Shashi plays tennis with Ravi (Deep weight = 2). Ram also knows Nandini who he met at a conference (Deep weight =1). Both Nandini and Ravi work at ‘ABC Corp’ regarding which Ram needs information.

Let us just say Ram searches for ‘ABC Corp’ on LinkedIn. A smart algorithm which takes interactivity into context will evaluate deep weights of both connection trails. In Nandini’s case it is simply 1. But for Ravi the algorithm can add the two interaction weights i.e. 5 and maybe apply some operation to account for the fact that Ravi is a second degree contact in Ram’s network, say divide. The evaluative deep weight now becomes 2.5 which is still greater than 1. The Algorithm will hence suggest Ravi over Nandini as a better connection to achieve a task. We will need much more sophisticated algorithms for actually evaluating ‘deep weights’ and evaluating various degrees of connections by using these deep weights.

However, this should have a fixed scale and constraints. Ram asking Shashi to talk to Ravi who then introduces Ram to to Balvinder who then introduces Ram to Sheetal just doesn’t make sense. Rarely will we find such high degrees in connections, but we need to account for this. The current system of degree of connection has fixed step values -1^st,2^nd,3^rd but maybe adding the context as above may implement a completely new ranking in the background with free flowing values.

A case can also be made to look at and derive data from other sources like Quora, Yammer or measures of interaction at a workplace say Exchange Information. These ideas seem good to ponder on too but the concept remains the same.

Finally -

• Just use the data in the background. That’s it - As interactions on LinkedIn and other social networks like Facebook and Twitter are inherently different, I have my doubts on more UI based Integration amongst these networks

• Make this a premium feature - The very nature of data that will be crunched and the advantage that will be provided screams it out to be a premium feature. It provides an edge and the edge should come for a cost. Also this exclusivity should make users comfortable in allowing LinkedIn to use their data

• Not just results, facilitate communication - For such a functionality, it is also critical to provide a way to get introduced via somebody in more seamless way. Adding intermediate connections automatically to CC automatically provide contextual information within the messages, go figure!

Finally, I would like to end with something I read-up on twitter very recently –

‘Sauthi Moti khan – olkhan’

(Guj for: ‘NetWork is Net Worth’)