6 Exercise: Project Progress - Video Tutorials & Practice Problems
Video duration:
6m
Play a video:
<v ->All right.</v> It's time to get back to your project. How are you going to solve the digital marketing analytics problem that you chose. So let's go back to our strategy framework and let's look at the next step, which is to assess what kind of obstacles could stop you. What could get in the way of you being able to fulfill that vision that you have. And some questions for you to ask yourself are, "Why hasn't this problem been solved already?", "What's gone wrong with similar projects in the past in this organization?" And you can try a technique called a pre-mortem. A pre-mortem is, if you think about putting yourself into the future, after the project is over, and you tell yourself, "The project failed, it did not succeed, we had obstacles we couldn't overcome." Now, I want you to use your imagination and say, "If the project failed, what do you suspect went wrong?" "What happened that you didn't plan for?" "What happened that you didn't have an appropriate plan for?" And use your imagination of putting yourself in the future and conduct that pre-mortem to decide what are the things that you're the most worried about that could happen. Those are the things that probably should fill up your obstacle list. As you do it, don't forget the power of why. Don't let yourself sit with surface observations, that aren't terribly actionable. Because the first answers to your questions, often feel like you're describing some type of state of the universe that cannot be changed. So in this situation, on the screen here, it says, "We don't know the right target segments." Well, why don't you? "Well, we don't have the customer data." Well, if you just stop there, that just feels like some law of the universe, and there's nothing you can do about it. But if you use the power of why, you can start to cycle your way through additional layers and get to the point that you say, "Hey, the problem is, we're not providing enough value to our customers for them to fill out our surveys. And so what can we do about that?" That's what your plan should be about, because that's what the real problem is. So don't forget the power of why as you're going through these exercises. Next, I want you to decide, what kinds of A/B tests should you run? What kind of criteria would you use to even decide what A/B test to run? So should it be something that's high value, it'll make a big business impact, maybe something high traffic, maybe it's a page or a journey a lot of people use. Make sense that it might be something that's low performance, something that's not working all that well now. Maybe something that's easy to change. Something where you understand the audience really well. And you know exactly how you're gonna measure success. But you could also pick something that doesn't have any of these characteristics, but maybe, it has some strategic importance. Maybe if you could get this right, maybe it would start to become high value, would start to become high traffic. And so you can use all sorts of different means to test whether something is a good candidate for conversion rate optimization, for your A/B testing. What I want you to do is to think about what the criteria is that you wanna use to select candidates for A/B testing. Next, what I want you to do, is to think of some very specific A/B tests that you should run, in order to help solve the marketing analytics problem that you chose. Be very specific. What are the ideas in your plan that you're least sure of, those are the things to test. So which page or journey in the experience, what is your theory about what's wrong and how to fix it, do you have any historical data that gives you some research that might clue you in to what the problem is, or is this really to develop that research, are you gonna learn about the problem by doing some A/B testing. You can also cheat and look and see what your competitors are doing. They probably have similar customer experiences that you do. And so take a look at what they're doing and see if you can find some ideas there, that you think might be worth testing for your site. So figure out what ideas you're gonna try and come up with several possible ideas, and maybe prioritize one of them, as the first one to test. Last, I want you to develop a lead scoring algorithm. It can be simple. It can look just like the one that we showed you for our client. So you can use explicit data, you can use implicit data, and you can also use negative data. So don't skimp on any of those factors. Now, you might be wondering, why would I try to come up with this basically dumb lead scoring algorithm, when I could do research, or use an AI model, or get some data scientists to help me come up with something that's really good. Well, you can, but come up with this one first. Because even if you end up doing the more complex one, these factors that you come up with will be the first set of things that your data scientists would ask for. They're gonna ask for what your theory is for what a good leads lead would look like. And so you might as well do this and come up with it, and you can put it to work while you're waiting for the data science team to go through all the effort that they would need to do, in order to do something that's even better, than your simple lead scoring algorithm that you come up here. But think about whether this is something that you could use, as you put together the solution for your marketing analytics problem. And if it is, put together a lead scoring mechanism that features the characteristics that you think make the biggest difference. So let's summarize what we have to do. So you're gonna continue to work on your plan, by first identifying the obstacles that could get in the way and derail that plan for your vision of solving your marketing analytics problem. Next, we want you to choose a few ideas for A/B testing. Think about what the criteria is, for how you're gonna select, and then, go ahead and pick a couple of possible tests, and prioritize at least one to do right away. Last, we want you to look at a lead scoring algorithm. So think about what the characteristics are that you think best identify the leads that sales should be prioritizing first. And figure out how you're gonna use that customer data to do it. All right. Go get them.