 This is a 20 minute session, so I will try and keep it short and sweet. But if you have questions, do stop and ask, okay? So I'm going to start off by telling you guys that this is based on our real experience. This is not being tested. This particular model that we're going to speak of is not being tested in the open a lot. We just used it at our own company at GeekTrust, and it found out it worked for us. So we're sharing this with you. Lots of things we still don't know yet. We don't know how we'll do that. We don't know how we'll do it in a complex team structure. But this is what we learned, and we have to share that, all right? Now, before we start, and we'll start with introductions, just want to know how many, say, developers in the room, okay, product owners, Chrome masters, okay, and project managers, testers, cool. So we've got a good mix, right? And I think this will impact almost everybody. Anybody working at a startup? Anybody working at a startup? Nobody? Okay, one. All right, so you might understand our pain better than the others, but let's see, we'll try and explain what we're trying to do. So quick introduction. My name is Krishnan. I used to work at ThoughtWorks. I worked at ThoughtWorks for 10 years. I was mostly doing project management. When I left ThoughtWorks, I was handling delivery for the entire Bangalore office. So I have a five degree of experience in agile, scrum, coaching. I used to do coaching in the UK, in Australia, in the US, and back here in Bangalore as well. So my primary role has been in ThoughtWorks and in Agile and all that. I'm Sneha, can you guys hear me? Yeah, okay, I'm Sneha. I also used to work at ThoughtWorks. I was at ThoughtWorks for about six years. I started off as an application developer. I've played various roles, moved into business analysis and project management mostly. And earlier last year we started up with Geek Trust. We're in the recruitment space. You can read more about us later. We don't have time to talk about that now. Now, moving on, how many of you have had the opportunity or the need to prioritize features? Hands up please. How do you prioritize? Today, how do you prioritize? What's that? Value. Value to the client. Who else said anything else? Gut feel, good answer. Okay, so those who said business value, how do you define that business value? How do you go about getting that business value out? A client says, yeah, I want this and I want it like yesterday. How do you define the business value for that? Against the thousands of other user stories that you already have. I mean, we've been doing agile for, or rather we have been agile for more than 10, 15 years now. How do you do that? How do you get to a business value for a story and say, hey, this gives me more business value than that? Yeah, okay, okay, excellent. So that's one way of doing it. For that, I'm going to give you a chocolate. Good catch. Yes. So if you have 50 different options or in a real world scenario, you have 100 different things. You can't push all 100 out, right? How do you select the four you choose for A-B testing? What do you do? What's the? Excuse me. Fair point. There you go. Okay. So that's the problem. If you notice, right, that seems to be an area which agile does not or rather I personally have not seen a lot of real world examples that work. I mean, I've run sessions for clients between Sneha and I. We have run multiple sessions for clients, maybe about 40 or 50. But when it got down to our own world, right, of actually spending money from my pocket and prioritization, I didn't want to do a poker game and prioritize what's important for me. I mean, we're a startup, there are four, five of us and all of us have the same say. It's not that Sneha owns product and I own development as my startup friend can attest to that. But everybody's say is important. Now, three of us are debating which should go live, which feature we should pick first. Whoever shouts the louder should not win. That's not the way it should be. So that's the problem we were looking at, right? When we started out with Geektrust, we found out that this was a problem, right? We didn't know how we would prioritize what. It's always like Anand was saying, he got a chocolate and ran away. As Anand was saying, it's gut feel. And gut feel just was good enough. Let's pass the problem of prioritization for a bridge and at Geektrust, like as we said, we are a startup. So what, we are a bootstrap startup. So the most important thing for us to run our business is our revenues coming in. And if you look at how we get our revenues in, how many of you have heard of a sales funnel? And how many sales folks in the room? Anybody in sales? Okay. Okay, okay. So sales funnel for us at Geektrust is nothing but a funnel through which our customers go through the different steps through which our customers go through and the end of it we make money. That's sales funnel for us. And for example, so Geektrust is an online recruitment platform. So at a very high level, any user who comes onto our platform registers, goes through a few things. And finally, if they get placed at a job, that's how we make money. So for us, the sales funnel is the different steps that a user goes through, which ultimately leads to revenues for us. Is that clear? And I think that's simple enough. Like if you pick even an e-commerce player like Flipkart, that's how they make money. That you have somebody landing on the page, now you take them through the entire process where they buy something and they get money, they get a commission. Probably not, but let's imagine they do. Right? So that's the situation. Is that clear? So sales funnel is clear and the problem state of prioritization is clear. Now, we were drawing out the sales funnel for Geektrust, right? We were looking at what are the different stages and then it just struck us. One day we had a realization. What if we married the sales funnel with our product prioritization? Right? Because for us, what is the most important thing? We are about to start by choice, by the way. What is the most important thing to make revenue? To move a customer through the sales funnel. Ta-da! Ta-da! That is how we make money. So we came up with this model. This is a very, very poorly drawn funnel. I admit that I drew it myself. Sneha has better handwriting than me. But this is how we came up with this idea of funnel-based product development. You prioritize your features depending on which part of the sales funnel it hits and how you want to move your customer forward. Sort of way, I mean, the idea makes sort of sense. We'll go through the detail. I'll pick an example and we'll go through that. But any questions so far? Is this idea sort of, does it, intuitively does it make sense? Yeah, especially a B2C player, I mean B2C company which is, you know, the product is the website. Yep. So even for you to figure out the first journey down, which is your MVP, you could even define the whole funnel itself. Like, initially you may not have the full funnel built up. If you think of the funnel as exact step, you might have, instead of say six steps, you might have four steps. So it still applies to your MVP itself. And even for your MVP, you'll still tie it down to what is it that gets you revenue? Or what is it that gets your customer converted on your site? Imagine you are starting up tomorrow. Which part of the funnel will you hit? So I'm not sure about that. I'm not sure I agree with that. And because as a startup and what we want to do is, I'm not worried, and this is from, this is what he really did, okay? I'm not worried about this part of the funnel. For me as a startup in my first two months, I'm just worried about this. Exactly, exactly. I want people to hit my bloody website first. Right, and get their email addresses. So the first thing we did was just a page which said, hey, we are coming soon. Are you interested? So now what we'll do is, we'll use a real world example, not a real world example, but an example. Imagine this website called smellniceonceinavile.com. Okay, it's a new website that we're coming up. It helps you buy perfumes online, okay? So that's our example. That's a website all of us are going to start up with tomorrow, right? It's a pretty big team. There are only a few developers. There's the operational heard, we'll lay off tomorrow. But imagine we are starting up. I'm joining, I'm overhead myself. So for our website, smellniceonceinavile.com, we need users to come and buy perfumes on our site. So that is the business model. That's how we make money. So before defining the funnel for this website, let's figure out what are the profiles? What are the users? What are the potential user base? And that's where user persona comes into picture. Do you guys, have you guys heard of user personas before? Okay, who's not? It's okay to say you're not, yeah. Okay, fine. So user persona is nothing but, imagine you create a, you profile your user base and you pick the one, the pick the few user personas that are most likely to come to your site. And a user persona is nothing but defining the whole life of that person in terms of what does the person do? What does the person like? Like, just like? Yeah, and also where is the person in terms of how, well he understands, he or she understands your business. So for example, you could have people who don't understand e-commerce at all, who are very new to e-commerce and that will be one set of user personas, whereas on the other extreme, you'll have somebody who's shopping on an e-commerce site day in and day out. So just profiling your users and then, what you end up doing is pick up that one person that covers most of your user base. And in our case, let's think of a persona, say Armani Anand. So Armani Anand is a persona who's very aware of the perfumes that he uses, who knows what perfume brands exist and who buys perfume online. So he's on the other, he's on the extreme where he understands e-commerce, he understands his perfumes really well and he knows what he's buying on your site. So for those who know user personas, does this make sense or not? Imagine Armani being a high end perfume you wanna buy and this guy is called Anand. So it sort of naturally clicks. This is my main user. And now we apply the user persona to the funnel, to the sales funnel that we had earlier, which is again in my poorly written handwriting, Armani Anand at the beginning of the funnel, we have done something to get him to the beginning of the funnel, then he searches for what perfume he wants, he sees the reviews, he adds to cart, et cetera. I mean, you know how it is, this is just a sample, but he goes or she, in this case, he goes through the funnel, right? This makes sense, any questions? Armani Anand plans and he goes through the funnel and when he finally makes a purchase, is when we earn revenue, right? What do you guys think and girls think? What's the next thing we should do? Now we have Armani Anand going through the funnel. How does this help with prioritization at all? So the next thing we did is, yeah. So the next thing we did is we put conversion numbers. So basically tracking Armani Anand through the site or the Armani Anand similar personas through the site, what is the conversion rate? How many people who search on the site actually end up seeing reviews? How many people who see reviews end up adding the product to cart? And so on, all the way up to the end of the funnel. So these are actual conversion numbers that every business should be tracking because if you don't know what your customer's doing on your site, then you have a big problem. Yeah, yes, sure, agreed. Yes, so it may not, see in your particular case, the funnel may be a little different. Exactly. Right, the funnel may not have a search in it. Right, it'll just have four products in it and say move forward. It could be different. And to think of a very, very MVP thing, the funnel might just be a search and buy. That's pretty much it. Exactly, so the funnel is just an, exactly, exactly. So this is just an example, but your steps in the funnel could vary based on where you are with your product, how early, how late in the market you are and a lot of other things. And I'd like to also talk about how important the conversion numbers are. So we are very small, our numbers are not high, the number of people hitting is not high, but it gives us patterns. It shows us that out of X people hitting our site, 3% convert. Out of 3% who convert, 4% update their profile. Out of that, the next percentage is this. It's not too difficult to do. Nowadays with Google Analytics, with Mixpanel, with all of these things, you can easily find this out, right? And this is extremely important. You need to know what are the stages and what are the conversions to the different stages. Was there a question then? Yeah. This is just a sample channel, but great question. Bad catch. But yes, you know, as it gets more complex, you could have multiple funnels, right? And some funnels won't be a straight funnel. There might be a feedback loop in the funnel. It could be different things that could happen, but this is just one channel. I'm just picking a, what's it called, a strawman. And based on different journeys, also you could have different funnels. They may not be just one way to end up buying a product on your site. There could be different ways. So there'll be different funnels for different things that you could do on the site. Yeah, and this is, yeah, I'll talk. How do you prioritize across the journey, right? You have a customer journey which is there. What we are trying to show you is, how do you decide which part of the journey you're going to pick to build first? Yes. Yeah, conversion numbers is the most important thing in this funnel-based product development. Yep. Move to the next stage. Just move, yeah. Yeah. Yeah. So this funnel could have different branches also based on what you could do and they could directly end up there as well. Right, see, like, I mean, I tried to keep the numbers sort of realistic, like 40% of people see reviews and probably 2% had to cut. But once you get them to the payment page, bloody well, 80%, 90% should move to the next stage, right, that's what you would expect. Yep, yeah, good catch. I was hoping you wouldn't ask this question. This may not work in that context. So if your product is not your business, right, if your product does not lead to direct sales for you, this may not work. But it still applies because you're still building something, even if it's B2B. I got one for you as well. Exactly, yeah. Yeah, so for the enterprise product, if one of your sales channel is through the web, is through online sales, right, you could have a landing page and you could use that, right, you could see just customer acquisition could be a funnel, right, people who register and say I'm interested, that's it. Yeah, see, it could be, but even if it's not cloud, right, you still need every single company in the world has a website and it says, hey, contact me for more details, that is also a funnel, somebody lands there, you pick that curiosity, you give them information and they say, I'm interested in this product, tell me more details. It's not the same as if this is much, much more effective in a B2C sort of environment, but in B2B also it works. And cloud, yeah, you're right, absolutely it works. Even in the B2B case, right, so here it's easy for us to pick that, the most important thing for this business is revenue. In a B2B case, it may not be that, the most important thing for you is, are people using my product? Exactly, I was just coming to that. So I'm saying, you don't necessarily need to tie to revenue, it's a B2B where it's not direct revenue, it could be how engaged are people, how people are using your product. See, you could do this for Twitter as well, right? In Twitter, the goal of a Twitter product manager is more engagement. So the output will not be revenue, it will be engagement, but there are different parts of the funnel. See, that, actually let me park that. I think in that B2B context it doesn't, what you said just now doesn't work, because it's not in that context that we are saying here. This is purely from a B2C model. B2B it'll work, but for the example you quoted this won't work. So I'm just going to move a little bit because we have like two minutes left. I'm free? Yeah. And then, you know, you wanna put your target numbers, right? You put your target numbers saying that, you know, initially you had 20%, now you want 30%, now you want 40%, then you wanna move it to 50 from two to five. Okay, let's pause here. This is where it is really useful. Now, every time Sneha tells me, let's build this feature, you know what question I ask? I ask her, which part of the funnel does it hit? And how much will it increase my conversion rate? The conversations are all objective, there is no gut feel, and you are forced to take a stance, right? If Sneha comes and says, you know what, let's build this cool feature over here, and I'll go like, but our conversion that is already okay, you don't have to worry about it. Let's pick this part of the funnel, which feature? So every single feature at Geek Press today, we start by saying, which part of the funnel will we hit with this feature? And how much do we expect to increase the conversion by? We didn't do any poker planning, this is very simple to do. And, you know, as a startup, you decide and prioritize features every two days, every three days, and we always go back to this, which part of the funnel does it hit? How much will it improve our conversion by? Was there a question at the back? See, if you're tracking your customers, right? You'll know if your funnel is right or wrong. You'll notice that customers go through a certain path, and if your conversion numbers are not adding up, you know that something is wrong. There are a lot of analytics tools which help you out with what your customers are doing on your site as well, and that itself is a validation of what where your funnel is headed to. You start with the hypothesis, then you test it on the real world. It would not? It could be, it could be, yeah. You'll need to figure that out. Yes, question at the back. That is true, we know our current numbers, that is factual, right? It is 65%, right? Target numbers, through agile mode, we try and make it incremental. We don't go for big leaps. From 60%, let's make it 65%, or 63%. We can't afford to make big leaps as well. So that's how we do it. So that's pretty much it. We just want to introduce this idea that we had. It worked really well for us. Not sure if I was at ThoughtWorks whether I would be able to implement this, but for a start up, this works out beautifully. We don't have to fight as much anymore. We'll just say which part of the funnel do we hit with this feature? And why is it more important than other features? Yep, yeah. I think that, of course, that plays into it, right? If there is something which is very small and easy to do. Actually, I'm not so sure. You know, we have two developers in our team, right? Even if it's very small and easy to do, even if it takes only four hours to do it, I would rather not do that, right? Do the thing which hits the funnel the most and increase the conversion number. If we had the time to, we would definitely do that. Yeah, we should. Quite honestly, we've not started that yet, but it's on our to-do list of doing AB at least two rounds of the thing. Right, right. So we see whether there's an uptake, depending on the, it depends on the feature, quite honestly, right? If it's something on the, beginning of the funnel, right, it's easy to test. We'll know within a week whether it works or not. We built a feature, our target was 10,000 hits, we hit 2,500, it didn't work. It didn't work as much as we wanted to. So it depends on feature to feature, but within a week, you'll get a sense of whether it's starting to move up or not. Yes, yes. Course correction is like every minute, especially at a startup. Yeah, yeah, yeah, yeah. Exactly. Cool, that's our talk, that's all it is. Thanks a lot.