Should you have an analytics team?

In an earlier post a couple of weeks back, I had talked about the importance of business people knowing numbers and numbers people knowing business, and had put in a small advertisement for my consulting services by mentioning that I know both business and numbers and work at their cusp. In this post, I take that further and analyze if it makes sense to have a dedicated analytics team.

Following the data boom, most companies have decided (rightly) that they need to do something to take advantage of all the data that they have and have created dedicated analytics teams. These teams, normally staffed with people from a quantitative or statistical background, with perhaps a few MBAs, is in charge of taking care of all the data the company has along with doing some rudimentary analysis. The question is if having such dedicated teams is effective or if it is better to have numbers-enabled people across the firm.

Having an analytics team makes sense from the point of view of economies of scale. People who are conversant with numbers are hard to come by, and when you find some, it makes sense to put them together and get them to work exclusively on numerical problems. That also ensures collaboration and knowledge sharing and that can have positive externalities.

Then, there is the data aspect. Anyone doing business analytics within a firm needs access to data from all over the firm, and if the firm doesn’t have a centralized data warehouse which houses all its data, one task of each analytics person would be to get together the data that they need for their analysis. Here again, the economies of scale of having an integrated analytics team work. The job of putting together data from multiple parts of the firm is not solved multiple times, and thus the analysts can spend more time on analyzing rather than collecting data.

So far so good. However, writing a while back I had explained that investment banks’ policies of having exclusive quant teams have doomed them to long-term failure. My contention there (including an insider view) was that an exclusive quant team whose only job is to model and which doesn’t have a view of the market can quickly get insular, and can lead to groupthink. People are more likely to solve for problems as defined by their models rather than problems posed by the market. This, I had mentioned can soon lead to a disconnect between the bank’s models and the markets, and ultimately lead to trading losses.

Extending that argument, it works the same way with non-banking firms as well. When you put together a group of numbers people and call them the analytics group, and only give them the job of building models rather than looking at actual business issues, they are likely to get similarly insular and opaque. While initially they might do well, soon they start getting disconnected from the actual business the firm is doing, and soon fall in love with their models. Soon, like the quants at big investment banks, they too will start solving for their models rather than for the actual business, and that prevents the rest of the firm from getting the best out of them.

Then there is the jargon. You say “I fitted a multinomial logistic regression and it gave me a p-value of 0.05 so this model is correct”, the business manager without much clue of numbers can be bulldozed into submission. By talking a language which most of the firm understands you are obscuring yourself, which leads to two responses from the rest. Either they deem the analytics team to be incapable (since they fail to talk the language of business, in which case the purpose of existence of the analytics team may be lost), or they assume the analytics team to be fundamentally superior (thanks to the obscurity in the language), in which case there is the risk of incorrect and possibly inappropriate models being adopted.

I can think of several solutions for this – but irrespective of what solution you ultimately adopt –  whether you go completely centralized or completely distributed or a hybrid like above – the key step in getting the best out of your analytics is to have your senior and senior-middle management team conversant with numbers. By that I don’t mean that they all go for a course in statistics. What I mean is that your middle and senior management should know how to solve problems using numbers. When they see data, they should have the ability to ask the right kind of questions. Irrespective of how the analytics team is placed, as long as you ask them the right kind of questions, you are likely to benefit from their work (assuming basic levels of competence of course). This way, they can remain conversant with the analytics people, and a middle ground can be established so that insights from numbers can actually flow into business.

So here is the plug for this post – shortly I’ll be launching short (1-day) workshops for middle and senior level managers in analytics. Keep watching this space 🙂


JEE Results

Exactly ten years ago, they used to give a sum total of 3400 ranks for IIT-JEE. Typically, to get an engineering branch at one of the “big 5” IITs you needed to be in the early 2000s or better. Back then, there were ~40 people from Bangalore who made it to the merit list (I’ve forgotten the exact numbers but if I remember right, at least 30 people from Bangalore JOINED some IIT or the other). About 1.2% of all successful candidates back then were from Karnataka (for IIT/JEE purposes Bangalore = Karnataka since there are no other centres in the state).

JEE results for this year came out yesterday. Most of the second page of today’s The New Indian Express is spent in giving footage to people from Bangalore who got a rank. This year, they gave out 13,100 ranks, of which 58 were from Bangalore – 0.5% of all successful candidates. And you have the New Indian Express which puts the headline “City Students crack IIT by the dozen”. Yeah, five dozen out of thirteen kilopeople is worse than three dozen out of three kilopeople. But anyway…

Back in my days, there was one decently established factory and a couple of fledgling factories in Bangalore. The established factory (a small scale industry by national standards) had 100 students, of which over 30 got ranks in the JEE (and about 20 actually joined IIT). Today the same factory has some 500 students. And surely not more than 58 of its students could have cleared the JEE! And then there are several other factories in the city. Don’t know if any of them have done significantly well.

Madness. Sheer madness. I had written about this before.

Postscript: I must admit there is a small bit of hotteuri (stomach burn) at the amount of footage toppers get nowadays. Back then, it was an advertisement by the coaching factory in all major English dailies in the city, and little else.

Postscript2: This post might sound like one old thatha sitting in his armchair and ranting. It is meant to be that way.

Arranged Scissors 5 – Finding the Right Exchange

If you look at my IIMB grade card, one subject stands out. It is one of the two Cs that I have on the card, and the other was in a “dead rubber” (5th/6th term where grades didn’t matter for placements). This C was in introductory marketing management. Where the major compoenent was a group project called the application exercise (ap-ex). I frequently crib that I did badly in that project because four out of six people in my group did no work, or even negative work (and this is true). Digging deeper, however, I think the more fundamental issue was that the two of us who worked didn’t really know what we were doing. We failed to understand the concept of STP till a few years after the project was over.

STP is one of the most fundamental concepts in marketing. It stands for Segmentation, Targeting and Positioning. I quickly appreciated Positioning, but took a long time in trying to figure out the difference between segmentation and targeting. In my defence, they are highly inter-related concepts, and unless you look at it from the point of view of social sciences (where each unique point fetches you one mark in the board exam) it is not intuitive that they are separate concepts.

So you segment the “population” based on various axes. Taking these axes in conjunction, you end up “segmenting” the population into a large number of hypercubes. Then you do the “targeting”. Find the set of hypercubes that you want to sell your product to (in the context this post is about, sell yourself to). And so once you have found your “target segment” or set of “target segments” you “position yourself” and go out to sell. And then you need to figure out the “4 Ps” of marketing. Product (fixed here – it’s you). Price (irrelevant if you don’t plan to take dowry). Forgot one P. The other is Place (where you will sell).

The arranged marriage market can be broadly be divided into two – OTC and exchanges. OTC (over the counter) is the case where you have a mutual acquaintance setting you up with a counterparty. The only difference here between arranged and normal scissors is that in the arranged case, it is your parents who are set up with the counterparty’s parents rather you getting set up directly. Since it is a mutual acquaintance doing the setting up, the counterparty is at max two degrees away, and this makes the due diligence process a lot easier. Also, you have one interested third party who will keep nudging you and pushing hte process back and forth and generally catalyzing it. So people in general prefer it. Historically, there were no formal exchanges (apart from say a few “well known village elders”). Most transactions were OTC.

One problem in financial OTC markets is counterparty risk (which is what has prompted the US government to prop up AIG) but this is not a unique problem with OTC arranged marriage market – counterparty risk will always be there irrespective of the method in which the relationship was formed. Apart from providing counterparty protection, one important role that financial exchanges play is to improve liquidity in the market. The number of transactions that happen in the exchange ensure that the market is efficient and prices are fair. Liquidity is an important asset in the arranged marriage exchanges also.

The problem that I’m trying to describe in this post is about segmenting the exchanges based on their most popular commodity types. I don’t have reall live examples of this, but then for each product you will want to go to a different exchange. For example (this example may not be factually correct) both the Chicago Board of Trade (CBoT) and Chicago Mercantile Exchange (CME) trade in both corn futures and cattle futures. However, the volume of corn futures that are traded on CBoT is significantly larger than the volume of corn futures traded on the CME. And the volume of cattle futures traded on the CME might be siginicantly larger than the corresponding volume in CBoT.

So if you want to buy cattle futures, you are better off going to the CME rather than the CBoT since the former has significantly greater liquidity in this product, and thus you are assured of getting a “fairer” price. Similarly, to buy corn you should rather go to CBoT than CME. I suppose you get the drift. Now, the same is true with the arranged marriage market also. If you want to get listed on an exchange, you will need to make sure that you get listed on the right exchange – the exchange where you are most likely to find people belonging to your target segment.

To take an example, if you think you want a Tamil-speaking spouse, you are significantly better off listing on rather than listing on, right? Of course this is just a simplistic example which I have presented because the segmentation and difference in markets is clear. Things in the real world are not so easy.

There are various kinds of marriage exchanges around. In fact, this has been a flourishing profession for a large number of years, and even the recent boom in louvvu marriages has done nothing to stem the flow of this market. You will have every swamiji in every mutt who will want to perform social service by opening a marriage exchange. Then, you have a few offline for-profit exchanges. Some of them work on a per-deal basis. Others charge you for listing, since it is tough for them to track the relationships that they’ve managed to create. Then, this is one business which has clearly survived the dotcom bust of 2001-02. The fact that this business is flourishing can be seen on the left sidebar of this page where I suppose a large number of them will be advertising. In fact, I encourage you to click through them since that will result in precious adsense revenue for me.

There is nothing wrong in carpet bombing, but that comes at a price. Notwithstanding the listing fees (which are usually nominal), you will have to deal with a significantly large number of “obviously misfit” CVs and bump them off. Especially if you live far away from the exchanges and have someone else broking for you, you don’t want to burden them too much, right? So the problem is in doing your segmentation and targeting. And then researching the exchanges to find which exchange has most liquidity for products belonging to both your segment as well as your target segment. And get listed on them ratehr than wasting precious time, energy and money listing on exchanges that are unlikely to be useful.

Since I began this (extremely long) post with marketing fundaes, I should complete it with some more (which is irrelevant to the rest of this post). A standard process for advertising is AIDA (Awareness-Interest-Desire-Action). Typically for a relationship to “happen”, you need a minimum of D from at least one of the parties, and a minimum of I from the other party. The normal arranged marriage process, however, assumes that an I-I is a sufficient condition for a sufficient lifelong relationship, and don’t give enough time and space for people to check if D is there. Hence the disasters. Hence the tilt towards the CMPs.

Arranged Scissors 1 – The Common Minimum Programme

Arranged Scissors 2

Arranged Scissors 3 – Due Diligence

Arranged Scissors 4 – Dear Cesare

Why is Ten Sports sitting on so many rights?

I wanted to stay up last night. I wanted to stay up and watch the WI-Eng match till the very end. Waking up this morning and checking the scorecard, it seems like it was a really good match. And Fidel Edwards seems to have become a last-day-shutdown specialist. This is the second time this series he’s hung on. And he’d done so once before against India at ARG.

There was another reason I wanted to stay up last night. I wanted to watch Liverpool play Real Madrid. I woke up this morning and saw that it was an amazing game, too. Looking through the Guardian Football site (btw, Advani seems to be advertising heavily on that site; it’s a pity he never advertises here on my site) I noticed that Chelski-Juve was also a strong game, despite the result. Another reason I would’ve wanted to stay up last night. For the record, I slept at 12:10. Tea-time in the Test match, and before either of the football games had started.

Ten Sports seems to have bitten off more than it can chew. It seems to own the rights to telecast too many different things. I think I have raised this point once earlier, but it pzzles me as to what Ten Sports is trying to achieve by getting rights to telecast so many things, most of which are happening at the same time. For example, over the last couple of weeks I’ve been unable to watch the first hour of WI-Eng even if I’d wanted to, because it was overlapping with the last hour of SA-Aus, which was being telecast at the same time.

The reason I slept off early last night was because I didn’t have the option to watch what I wanted. All the three games that I’d’ve been reasonably interested in were supposed to be on Ten Sports (Zee Sports doesn’t count since Tata Sky doesn’t offer that), and I  realized that I’d be forced to watch what the guys at the Taj Entertainment Network would want me to watch. Denied the option to choose what I wanted to watch, I went to bed.

It puzzles me that Ten Sports isn’t subletting its contracts. Devoid of anything decent to show, I suppose that ESPN or NEO would’ve only been too happy to acquire the rights to telecast last night’s Liv-Real game by paying a fee to Ten Sports. And it would’ve unlocked value at the hands of the remote-holder. Ten Sports need not let go of the rights to show all the games. All they need to do is to sell the “out of money options” – the rights to the game which they won’t be able to telecast anyway.

Now, the problem will be if accounting for all costs, no options are out of money. For example, you know you won’t be able to show Liv-Real. But you think that the loss of brand equity of your channel would exceed the money you’d gain by selling this option to another willing channel. The viewers are the only losers at this game, but I don’t know what can be done. After all, viewers  are way too dispersed in order for them to take any kind of action.

Extending this question, what can a sports body do to prevent a bidder from acquiring rights to telecast and then mess up the telecast (or not telecast it at all) ? After all, the sports body is out there to make as much money as possible from the TV rights, and they need to ensure significant investment into broadcasting by the broadcasters, so the “i’ll give rights to only those channels that are in the interest of the people” model won’t work.

One option would be to sell the rights to two channels in each market. But given that broadcast is a natural monopoly, the sports body will not be able to make as much by selling to two bidders as it can by selling to one bidder. Is there any other solution that you can think of? If yes, unleash.

A new paradigm for selling advertising slots

There are fundamentally two kinds of videos – videos for which willing to pay to see, and videos which you are paid to see. It is intuitive that advertisements fall in the latter model – for watching an advertisement, you are being “paid” a certain sum of virtual money which gets encashed when you watch the program along with with the advertisement appears.

You might also notice that despite all the hue and cry about copyrights and people getting videos pulled off youtube, it is unlikely to find a case where an advertisement has been pulled off youtube. An advertiser will only be too happy to have more people watching the advertisement, and by pulling it off youtube, the advertisor will be shooting himself in the foot.

When you are watching TV, and a painful ad comes along, you are likely to switch channels. Or get up and take a break. And turn your eyeball to the screen only when all the advertisements for that particular session are over. So, in effect, by showing a bad advertisement, a channel is reducing the number of eyeballs for the other advertisement in the same session (a session is defined as a consecutive set of advertisements, uninterrupted by the main program. it can run from approximately thirty seconds to five minutes)

On the other hand, a good, popular and well-made advertisement is unlikely to make the viewer switch channels, or get up. It is more likely to generate higher eyeballs for the other advertisements in the session – without any additional effort by the other advertisements in the slot. And thus pushes up the value added for all advertisers in that particular slot.

So the idea is simple – advertising slot providers (i.e. TV channels, etc.) should incentivise advertisers to make better advertisements. Or use the better advertisements more. And the simplest incentive you can give is monetary. So offer a discount for the better and more popular ads. So far, the model has been to make viewers view ads that come along with a programme. The new paradigm is to make viewers view ads because they are placed next to ads that viewers want to see.

I’m sure that once this kind of pricing gets implemented, it will be more profitable both for the TV Channel and for the viewers. TV Channels will be able to sell the “network value” of placing ads on their medium, and use that to more than compensate for the lost revenue in terms of discount. Viewers will like it because the bad ads will be gone, and they will be saved the trouble of switching channels each time there is an ad break.

There remains the small matter of implementation. We need a way for rating advertisements. Online/SMS polling will be no good as they can be rigged. Neither will youtube help. We will need to find a better way to gauge how much people in general find ads. If there is some way in which TRPs for ads can be measured, that would be helpful, too. I’ll think about this problem, and maybe publish a solution to it in due course. I urge you also to think about this kind model, and let me know if you can come up with any bright ideas.

One option would be for the channel to pick what it calls a “winner advertisement” and fix the various slots in which it is going to be played. Maybe the winner might be given the choice of picking which slots it wants to go in. Then, the channel can make the placement of these winner ads public to the other advertisers and encourage them to bid for the surrounding slots. This bidding can help gauge the popularity of the initial winner ad, and then the channel should share some part of the proceeds of the auction with the winner advertiser. And when the premium that other advertisers are willing to pay in order to get a slot close to the winner drops, the channel will know that it is not a winner anymore and replace it.

So what I have described here is some sort of effective peer-review process for advertisements. Different channels can choose different strategies for the order in which to let channels pick their slots, about what kind of auctions to hold, etc. The most important thing about this peer-review process is that here people are voting with their chequebooks – and when people do that, they are very likely to know what they are doing.

So think about this. I think it is a good idea, and it seems like one of those things that if one channel implements it, it will become some sort of an industry-wide standard. And if you are not doing this because you think you don’t have quantitatively inclines people,  the fired investment bankers are still around.