You're essentially trying to figure out sales all right cool so the regression the whole point of building a regression model is that you would understand the weightage which is nothing but the influence of each of these channels on driving sales right so that that sees I mean it's very against very simplified obscene or straightforward regression problem we would have to build it out in different forms but for the sake of understanding so why did we talk about attribution again yeah maybe the fact that the consumer the human psychologist - why do people buy I'm saying it's not as simple as what they teach you in a Philip Kotler textbook which is to build out your eye dance you know that somebody's awareness becomes action as long as you hit them with a brand campaign here.

A seal campaign here that is not the way we do it but jumping on so what is the difference between a strategy and an objective or a tactic and this is a fairly straightforward explanation I know why they would want to call something very simple as this as an answer matrix but ultimately it boils on the fact that these are four major strategies that any company would follow at any point of time right so your in the existing market and you're trying to sell existing products then you would essentially focus on trying to penetrate deeper into the same market as opposed to you are an existing market you try to come up with a new product you would focus your strategies to basically develop a new line of business or do product development for trying to sell an existing product into a new market you're essential would focus on developing that overall market for you right-size the pipe.

This is your riskiest strategy which is basically going into a new market with a new product otherwise known as diversification now I want you to sense a talk about I want to talk about an example as close to my heart which is something that's happening in my company right now so as I said Allstate is an insurance company so we're in the business of insuring cars and home properties right now auto as a market has been facing severe headwinds meaning we've essentially the overall auto market has been dwindling like crazy in the last three-four years so the premiums are going up and the claims are going up as well right so which basically means that we are we don't essentially have we are bleeding in that market.

Now so and where is the threat coming from so in the last couple of years so Google and uber have started testing our driverless cars so this is not a reality today but it could probably become a fairly big threat to us in the future so when you think about driverless cars right who do you actually go ahead and ensure the card itself okay good point so now the card itself as a consumer let's say they decide to start lending the car out to consumer right the consumer is not robbing he would be interested in trying to ensure the card yes because he wants to keep the car safe but that's more from our OEM point of view which is the manufacturer of my life. you right they would be interested in making sure their cars at the lease out or lent out are in perfect condition because I because I'm not sure how their business model is gonna play out.

When it comes to an end consumer a driverless car becoming owned by an end consumer that that's I I think the business model itself is still not too clear but the point is with this probably becoming a reality the insurance market for auto is definitely going to be threatened now our CEO whose name is Tom Wilson he came out with a vision statement saying by 2021 I want to make okay this is wrong this should be at least 2018 but it's actually a while our overall objective as a marketing team is to grow our enforce policies he's going into diversification me while we're traditionally we're in the auto market now we want to get into the live market so we don't sell life insurance l but it's a very small portfolio for from an overall line of business but the idea was let's get out of our toe or not let's get out.

Let's focus the while the size of the pie is decreasing and auto let's try and increase it by in another market which is life insurance that is the overall diversification strategy right he essentially said that at least 25% of all US households which are also our consumer would own an all-state life insurance policy so what does that mean for us as a marketer as you know on-the-ground analytics person so I'll talk a little about how my team is actually you know real life trying to realize that strategy right so okay do you guys understand the concept of a propensity model or the term propensity yeah penetration is probably like a post factor what happens after me propensity is nothing but you are the likelihood that you would do anything right.

So as a marketing analytics person I am charged this problem essentially I'm translating this overall vision statement into an analytics problem statement which is to say that I need to figure out that segment of my consumers you know population who already owns an auto policy but who has the highest propensity to go ahead and buy life from me right so that's the analytics problem statement so how to give sense he translated this into an analytics problem statement now how do you break the break up that problem statement which is to say identify all those guys who have already an auto policy and who would probably go ahead and buy you know the life so that becomes a propensity modeling problem so in the marketing analytics world propensity modeling is a fairly big term propensity modeling is nothing but you figure out who's your best target customer who has the propensity to buy your product as simple as that right.

So that's called propensity modeling but how do you make that a reality hopefully this should be interesting because Jason is talking about analytics problems in real life right so what we first did was so the US population is about 300 350 million-odd people right now in the US consumer information is a little more mature and more rampant as opposed to India India the consumer level the information is very sparse right we don't know enough about people because pretty much in u.s. you can track everybody's digital patterns right so what has happened in the US over the last few years is a lot of database marketing companies have sprung up and they've started selling data to you know companies to produce.

Experian is a fairly big database marketing company so all said as a company is actually bought data from Experian about close to 230 u.s. million folks which stand next about 100 to 500 25 US households so what we have in hand is 425 million households I know everything about their demographics I know everything about their credit history I know everything about the auto market whatever affinity and this is at you know a personal level this could be at a personal level as well or it could be aggregated at a zip level this is a zip love Limonov these terms looking at you you understand what's the demographic of any person right.

Post a Comment

Please do not enter ant spam link in the comment box

Previous Post Next Post