Product Marketing by Numbers - Objectives, Goals and KPI frameworks
In this talk Jono Alderson (Head of Insights at the fictional LinkDex) argues that data‑driven product and marketing work collapses without a clear definition of success. He describes how senior leaders often impose the "Highest Paid Person's Opinion" (HIPO), which can misguide teams away from consumer‑centric goals. Alderson stresses that before any tracking, teams must articulate objectives, break them into concrete goals, identify signals, and specify precise metrics (KPIs) with targets. He introduces and adapts several frameworks – the HEART model, a custom Objective‑Goal‑Metric template with bullet‑chart visualisation, and Pirate Metrics – to structure this process. He outlines a simple, low‑tech workshop (post‑its, CEO involvement) to surface stakeholder definitions of success, then refines them into measurable KPIs, ensuring they tie back to revenue, cost, or brand impact. The talk also covers the importance of metric specificity (e.g., qualified installs, composite success interactions), acknowledges tool and data gaps, and recommends using an analytics maturity model to assess organisational capability. Finally, Alderson lists practical tools (iPerceptions, Segmentio, Trello, Woopra) and a process for embedding tracking criteria into feature specs and user‑feedback loops, enabling decisions that are grounded in a pre‑defined notion of success rather than endless data rummaging.
Chapters
This is me. I'm Head of Insights at LinkDex, which is a made-up title at a made-up company. LinkDex is an enterprise SaaS data SEO platform, which is a whole combination of words which probably won't mean a vast amount. In reality, what I do is a combination of thinking about where digital marketing and SEO is going, and then feeding that back into our product roadmap and prototyping new ideas and new processes and tools. So I'm spending a lot of time thinking about what the users need and trying to interpret that into something that resembles product ideas and a roadmap. But before that, I was a bit of a hardcore technical SEO and analytics geek, and very much viewed those 2 things together as a single discipline. The idea that sophisticated, grown-up digital marketing should probably have some degree of data-driven thinking behind it. There's no point coming up with nice ideas for shiny campaigns and spending advertising budget if you're not in a position to understand what's happening off the back of that. But underneath all of that, I'm kind of secretly a bit of a developer, and I play with code, and I build widgets and websites, and I've got a few apps of my own. So I've kind of come through this backwards journey of coming from products to strategy and analysis, and now I'm finding increasingly that As certainly the product market matures, everybody is wanting to think about data and how we become more data-driven and establish processes to be consumer-centric and understand what people want and design our businesses around that. It's really exciting. But my concern isn't that we need data and we don't need shiny tools and we don't need to install tracking scripts and we don't need to understand what everybody does. Actually, what we need to understand is what we need to understand. And there's a fundamental crucial step that everybody misses, whether it's in product development, whether it's in digital marketing, whether it's in public sector, whether it's any of these scenarios, nobody stops and thinks, what does success look like? And how do we define our objectives and how we know if we're achieving those and the metrics that we use to understand our progress towards goals and forecasts and targets that we've set? Until you've got a plan, until you know what you're gonna measure and why and what you'll do with that information and what success looks like, there's no point counting anything. We get obsessed by the tools and the numbers and APIs and millions of rows of data and how we can segment and build cohorts and All of these wonderful things which are very, very valid, but only if you've got a plan and only if you know what you're gonna do with them. So there's a few challenges that make it quite difficult to get to this place where we've got a nice clear vision of what we wanna do and why we wanna do it and then go on to managing how we track that and make decisions. The first and foremost is the HIPO, which if you're not familiar with it as a term, is the highest paid person's opinion. This is the boss, it's the CEO, it's your line manager, it's their manager's manager. And their opinion is golden regardless of how ridiculous or unfounded or archaic it is. And if they say it's really important that we get loads of installs, or all that matters is money, or all that matters is our growth, when actually there are more complicated considerations like adoption and perception and user feedback and 1,001 other things, they become a bit of a trap. And actually the issue is their, their role and their psychology and the reason they are the people they are and the roles they're in is because they are defensive by nature. And their responsibility and their job is to question things and to worry about budgets and to challenge people spending money and thinking differently. It's absolutely fine and there's no way around that. But the challenge is they're not the people who are best qualified to understand what success looks like and to define that. You guys are. You guys are close to your product. You understand what your consumers want. You understand what they need. Your ability to think, how do we measure what success looks like, is vastly more sophisticated than theirs. But they shout louder and they have a bigger paycheck, so they define what good looks like, and people's behaviors start to align to that, and you miss out on opportunities to be truly consumer-centric. Which leads to a lack of direction. It means that you quickly go to your Google Analytics data, you quickly go to your cohorts and your big databases and your Hadoop, and you start looking in the hopes that you can find some kind of magic golden star of insight, some nugget that will tell you that one critical piece of information that you need to change your strategy to think about new things, that moment, that flash of light where you suddenly think, yes, if only we tweak this knob and this dial and change what we're doing a bit, suddenly we'll make millions of pounds and generate loads of installs. The thing is, it's not in there. There's no point going and rummaging and looking and trying to crunch through all these numbers and have 18 different levels of segmentation and cohorts in place, because until you know what you're looking for, you're never going to find it. And we all spend huge amounts of time rummaging through data in the hopes of finding these ways we can tweak our behavior, but it's never there. She's to analysis paralysis. There is too much, and you struggle to make any decisions at all, and suddenly we're relegated to being the people in the corner who have all the responsibility for getting it right but none of the tools to ensure that we can do it. I think this is the fault of the analytics industry. So this is Google Analytics' homepage, and I hate this woman with a passion. Apart from the fact that she has the most bizarre face on the planet, she lies. She lies. I can't quite read the text. I'm going to have to go over here and have a look. Essentially, it says, we installed Google Analytics into our product, into our website, into our app, and it showed us where to make changes, and we made thousands of percentages of increase. There is no button. There is no box that says what you should do. There is no, oh, and therefore you should. It doesn't exist. But all of the marketing collateral and all of the messaging of all the analytics and tracking tools uses this kind of information. It says, put this in, and you'll learn things, and it'll be great, and you'll know exactly what to do, and you'll make millions. There's a huge disconnect. It's a real issue. So I did my own cover on it, which I can't read the text, but it essentially says something along the lines of, we sat down over a cup of tea and thought about what success looked like and interviewed some customers and really got to the bottom of what we were trying to do. And then we ran it past our stakeholders and we got their buy-in. And then we installed some code and set some thresholds for what good looked like. And now when we hit those, we make decisions that we know what we're going to do with. and we're all really happy and making lots of money. That's what it should say. As a result of this, there's a real challenge with articulating what success looks like because you're put into this corner— where's my mic? There we go. Because you're kind of pushed into this corner by the hippo and this lack of definition of what good looks like, it becomes increasingly difficult over time to stick your hand up and say, what are we trying to do? What does success look like? Is that adoption rate good? Is it bad? Should it be better? I don't know. And it all just becomes a bit Bit flunky. And furthermore, you then start getting obsessive about comparing to the past, so you don't know what good looks like, so you start saying, well, last time we did this we had 100,000 and this time we've got 120,000. Yay, maybe, maybe that's good, not sure. And all targets start getting framed in the context of we need to do 18% better at this thing this time. And then when that doesn't work, you start looking at competitors and say, well, how are they doing? Maybe we ought to be 40% better than them. And the bit you're still missing is this idea of what are we trying to do and how do we know if we've done it? Skip through some of this, but essentially you end up in a position where your entire life looks like this, where you're staring at spreadsheets that get bigger and bigger because your managers and your senior people and the people you're reporting to aren't finding the nuggets of insight that they're wanting you to deliver. So it gets bigger and bigger, you get more charts, and everybody's unhappy. Sad kitten. So let's fix this. Let's fix this forever once. It's really simple, really, really, really simple. Essentially, the key is that you need bespoke data, and you have bespoke data, and you need bespoke processes. But there are really easy frameworks you can use to get to this. Who's come across the HART framework? A couple. So this is nice. It's cute. Google used this for running all of Google's internal stuff. And you get a nice little grid that says, for happiness, engagement, adoption, retention, and task success, HEART. Pretty universal, nice framework for any kind of producty stuff. What are our goals, what are our signals, and what are our metrics? And you fill it out, and you do something like, I can't read that. Our goals for engagement are for users to enjoy something, something, something, and keep going and retain something, something. That's really blurry. Signals, which are essentially the KPIs which say, how are we doing and how are we doing with this? And then the metrics, which is right down to the explicit definitions. It's quite nice, but it's a bit basic. You're not Google, so the same principle applies. You want something a bit more tailored and that can really answer off to a lot of those challenges from earlier. You want something like this that doesn't fit on the screen. Excellent. There'll be a link to these slides afterwards. Take this. This is the template. This is the solution to all of the problems I've just outlined. It's a universal process. It fits any product, any model, any business, any scenario, any campaign, and it allows you to control the conversation and to manage people's expectations and to set and define what success looks like. So in a nutshell, you can't quite see, but off the top left is the column for objectives, which says, what are we trying to achieve at the highest possible level? And then you break it down into goals. What does success look like in the context of what we're doing? So your CEO will have a very different perspective on what success looks like than you will in terms of your product. He thinks in terms of revenue and business growth and share price. You think in terms of adoption and retention. You then break that down to a specific metric and say, what is the KPI behind this? OK, well, it's the number of users performing this type of action per month on a rolling basis, in very explicit language. And then you can do things like acknowledge segments and cohorts and set targets. So I'll whiz through some of that. Bullet charts are the best graph in the world. These are great. You'll see them in here. The principle is really, really cool, that you need to contain how are we doing, is that good or bad, and what are we going to do about it, and do we need to panic, all in one little space. Any graph or chart you ever show will always lead to another question because it's not quite good enough. If you show a chart going up, you'll get asked, where's it going, where's it been, what were our targets? This conveys all that information really densely. So you've got a nice line saying we are currently at 80, our target was 75. And the colors of the background say, should we fire people or should we have champagne? Really, really straightforward. And if you set intelligent targets, you can visualize the whole of what are we trying to do and are we doing it and are we winning in a tiny, tiny space. And then you can essentially fit an entire understanding of your entire ecosystem on an A4 sheet. Really nice. So you need 5 things to do this. You need some pens, you need some Post-it notes, you need a wall, you need your CEO or equivalent, and you need some time. This is really easy. It's not intense. You need to get your hippo in a room, whether this is your boss or their boss or their boss's boss, and get them to articulate what success looks like to them. They will sometimes say share price. They will sometimes say happy users. They will sometimes be slightly more sophisticated and say task completion rate. Get all of that out of them. Mine it from them. You only get one chance to do this. If you try and do this over multiple stages, they get bored and disengage. You then need to remove the elephant from the room. So you need to get everybody involved in and who has ownership in your product success To write down the story of Hansel and Gretel in no more than 5 bullet points on a Post-it note without conferring. Get them to compare each other's and they'll be absolutely identical. They will say, some children, breadcrumb trail, witch, an oven, yay, and they'll be pretty much the same. Now these are people who don't talk about Hansel and Gretel every day, presumably. Get them to turn over the Post-it note and do the same for their business objectives. What are we trying to achieve? What are we trying to achieve? What does success look like? What is our business for? What is our product for? What do our consumers want? And you will get radically different answers. You will get mishmashes of mess, and it allows people to admit that they've been in a position where this has never been done, and it removes fault and blame. You can go through these, you can start to condense them and consolidate them into key themes, and essentially all you've got to do is make sure that any of them tie back to increasing revenue, decreasing cost, or improving the brand or the product or the experience. Anything that doesn't fit these is probably a KPI or a goal or something different, but anything that fits this model is, is your best chance at getting a solid definition of what success looks like. Then you need to start thinking about the KPIs behind these, and this is where it comes down to things like understanding that retention is your most important metric. And you need to be really specific because you need to remove all opportunity for wriggle room and interpretation. You need to make sure that your metrics are solid, that they tie back to your tracking capabilities, So the idea of number of conversions or number of installs is quite nice, but actually number of qualified installations is much better, because if 50% of people uninstall the next day, it's a pretty ridiculous metric to be making business decisions on. But actually even that's a bit flawed. You might want to think the number of them plus the number of people who've made valuable inquiries plus the number of valuable phone calls plus a whole bunch of other things that all constitute successful interactions. That's a metric that nobody can argue with, that will never get forgotten, that will be tied down to the point where nobody has the opportunity to argue or disagree with it, 'cause it's an explicit definition of what success looks like. Skip through some of this. The problem is you won't have the infrastructure to track all of these. You won't have the tools you need to do all of this from day one. You will need to explicitly acknowledge the things that you need to install new kit for, that you need to invest in new databases for, that you need to start tracking outright. And there's a really nice process for doing that, which if you acknowledge that these gaps exist at day one, it stops this process being shot down as a way of managing experiences. So this is really kind of the holy grail, and it applies to generally kind of larger, more established organizations. If you are smaller and a bit startup-y and a bit entrepreneurial, you might actually want to look at something different, which is Pirate Metrics, which is really nice. You can run these in parallel, but they're a slightly different way of thinking. This is quite cute. You select a handful of metrics from acquisition to activation to retention to revenue to referral, which shortens to ARR, hence pirate metrics. And you end up with something like this. So you've got a simplified funnel from how do people find us to how do we get them on board, how do we get them installing, how do we get them utilizing, to then How do we keep them? How do we retain them? What are the metrics around that? How do we get them referring and bringing other people in, and are there mechanisms for promoting that? And then how do we get them generating direct revenue? Interesting that that's at the bottom. If you chase that first, you don't find any app or any product or any business that starts off with the objective of generating revenue won't generate any revenue. This really is the output of those processes and getting it right. So, It's a really nice model. With all of this, it's quite difficult to know whether you're doing a good job of it and whether you're getting it right. So there's a slightly higher-level process that sits above this that's worth looking at. I've forgotten who produces this, but you'll— who? Stefan Hemmel of— not some— Cardinal Path, that's the one. Google— who? Independent. OK, but Google Cardinal Path Analytics Maturity Model, and you will find this, or slightly more more up-to-date, sexier version. And the principle is at an organizational and operational level, you wanna say how good you are and where your strengths and weaknesses are and recognize that. So look at things like, how good are we at managing our analytics ecosystem? How sophisticated are our tools? What's our methodology? Do we have a methodology? Is it one guy in the basement's methodology or is it a distributed team thinking? Or does it permeate the whole ecosystem? You can start to fill that out and you'll inevitably get something like this, which kinda says, you know, We're okay. And actually, our tools aren't great. We could use some more tools. And what we're really, really good at is thinking beyond just our products and maybe thinking about how that extends into the brand experience and our website and our culture. And we're not very good at the amount of resources we deploy. This gives you the permission to then start to think, how do we get better at this? Do we need more tools? Do we need more budget? Do we need more people, better people, different people? And you can get senior buy-in that you want to progress towards something like this that's says we're really well-rounded and we're really sophisticated. One thing worth noting is if you find that you've got really kind of disparate shapes, that's really indicative of a wider problem. You want to be growing this a little bit at a time in a really balanced shape. It's really nice for saying we've put loads of money into tools but we don't know what we're doing with them, or vice versa. So now you can start thinking about tools, and now you can start thinking about what clever data you want to collect and what you want to track. If you do this before this stage, you don't know what to do with it, you don't know what you're getting out of it, and you can't apply it to anything. This is my go-to call list. If there are any on here that are unfamiliar to you, definitely go and have a look. Start with— this side is kind of managing your ecosystem and this side is your data. iPerceptions in particular, really, really nice for products, which is about structured feedback mechanisms from your users. So things like getting metrics like task completion rate and satisfaction rate. Asking your people directly. What were you trying to do and were you successful and how was the experience? And then building cohorts out of that. Show me people we— show me users who sign up at a weekend who are dissatisfied because they can't find our support information. You can build these segments out in things like Google Analytics really easily and then you've got really actionable, really tactical stuff that ties back to your model. Segmentio is nice for managing your tags. Trello is a beautiful way of working and managing agile development processes. Wooper is really, really, really cool for tracking individual interactions over time. Go have a look. Really, if you've not seen and used any of these, go away and have a play. They are all beautiful in their own special way. A little bit on processes. What's that say? Build analysis criteria. So this is something I'm doing at Linqdax, is when we spec new features and when we are developing new ideas, We will have a comprehensive technical spec that says this is what it is and how it needs to work. Alongside that, we've started building documents that say this is how we'll track success. And when users perform this action, this is the type of data we'll collect. And it means that we're never rolling anything out that we can't understand whether it's working or not. And we'll say that 2 weeks after launch, we will determine this a success if these metrics have been met and if these criteria have been fulfilled. Similarly, you can start to fill in some of the gaps in that by encouraging and gamifying user feedback. So have little checkboxes and smiley faces or sad faces and angry faces that you allow people to click on to qualify their experience. If you've launched a new feature, you can see how people are using it, but you can't qualify the experience particularly easily. If you have a nice big smiley face that people are clicking on, you know that they're positive and they're engaged. Start to work out how you elicit that kind of feedback from people as part of that product spec. So let's bring it together. Really, this is about knowing what success looks like before you go out and start doing things and start tracking things. Otherwise, you'll never know and you'll just get deeper and deeper and buried in charts and analysis and metrics that don't really mean anything because you don't know what they mean. The senior people around you and above you are locked into your definition of what success looks like, which removes their ability to forget why you're doing a certain thing or to disagree with a strategy that they agreed with last week or to pull budget because they understand the connection between what they want to achieve and how you're going to measure and implement it. Anywhere where you can't do things in the short term, where you don't have the tools, you don't have the budget, rather than that being a convenient excuse for them to pull budget or forget about things, it's accommodated for because you can say, we know here we are, here's our maturity, here are the things we need, and we've got a roadmap to get there. And really all of this comes together to ensure that decisions are made about the things that matter and that you've already decided what the important things to you are, and there's no reason that you— you don't need to spend time rummaging around in analytics reports trying to find those nuggets. Thank you very much. Thank you.
