In this video we're going to talk about analytics. It's very important you're involved here as a product manager. Because as you begin to increment your existing product and you've figured out this existential product market-fit questions, the questions you have to answer are going to become more nuanced. And good, actionable analytics are really a interdisciplinary function of qualitative and quantitative and substantial understanding, as well as number crunching. You are the perfect person to make sure that that work, whether you're the primary person doing it or not, is driving to relevance and actionability. Now, the thing with analytics is, look at this number here, it is just so inherently convincing, right? And if I put a couple more decimal points, it would be even more convincing. We get fixated on the quantitative. We get fixated on the numbers. They're comfortable, we can just sit in our cube and play with them. But that's not where not good analytics come from, not exclusively anyway. This is just a sort of rough view of a way you might think about the analytics process in the kind of way that I think you need to as a product manager. We start with an analytics objective of some sort. Now, I am not assuming that this is particularly well formulated. It may come from some general idea you have, or it may come from some even vaguer idea that you get from management or a field person, sales person, account person. Let's go back to Cooped Up LLC. Let's say their CEO comes to the product manager and says, hey, I've been out in the field with some of the factory farms. And the farmers that are selling this one particular variety of chicken, the yellow delicious, they're experiencing incredible growth. And I just sold them a maintenance contract, because they were off of our maintenance contract. They're growing, they're ready to pay us to come in and do the maintenance for them. Go and help the sales people sell that to everybody else. And you would need to unpack then and figure out, okay, so is that really going to be true for all the other people that are growing these yellow delicious chickens? And do they want to buy maintenance? How would we test that? How do we even identify them? So the first step is to unpack that objective into an actionable set of steps and then to move to this thing of understanding the problem or the job to be done. And the key here is to begin with the end in mind. So what's the actionability of the analytics objective? And here, it is equipping your salespeople to go out and test this proposition that the yellow growth chicken farmers are experiencing a lot of growth and they would love to pay us to come in and do their maintenance so they can accelerate that growth. And maybe from experience, maybe from going out and talking to the people, the salespeople or the sales managers that are actually going to be responsible for taking this action, you know that if you just crunch a bunch of numbers and hand them over to sales, it's not actionable. It has to actually end up in the CRM as a set of call activities and things like that. Knowing this, you go and you diagnose the data you have. Can you identify farmers that have this yellow delicious variety from the data that you guys have, from the data that somebody else has? Maybe you should go out of your cubicle and talk to the salespeople, because maybe there's some proxy for the presence of these types of chickens that you can use that you weren't even aware of. And [COUGH] here, you generally want to figure out what data do we have, what data could we go go get, and what maybe non-obvious parts of the data could we use to drive to this action ability? Because you gotta be scrappy, you don't always have the perfect data sets. Then we need to figure out, can we get this done? And in what pace can we get it done? Is it going to be one big project that we need to completely finish? Or can we maybe iterate through this in little pieces and kind of test it in small batches? I won't tell you how important it is to do things iteratively in small batches, because I know I've been saying that, but it is. Then we prototype. So maybe we make a version of this thing that is just kind of something that we can show the salesperson or the sales manager and say, if we made it look like this, would it be actionable? And really, this is a placeholder for prototype and execute as you iterate through this. Execute, Then you communicate and observe this analysis to the person that is going to take the action and guess what? They may not be ready to really look at the data right when you show it to them. That's just the reality. Busy managers don't want to look at analysis until they want to look at it. So you may need to show it to them, introduce it to them, give some notes. Come back in a few days, a week, when they've actually encountered it and try to use it and see how things are going. And that is greatly helped, of course, by doing this in an iterative way and trying to move through this with observability about whether the analysis was actionable or not. The key thing is marry your qualitative and your quantitative understanding, and iterate a lot. And focus on the actionability of your analytics and what needle you want to move. How will we know if this is working? Will we be able to sell more maintenance contracts to these people? How am I going to close the loop on these analytics to see if we're really getting at the core of the objective? Those are some ideas about how to approach the analytics process. And I think it'll make your work around analytics with your collaborators much more actionable, much more relevant to your product.