From Static Documents to Persona AI: Making Customer Personas Actually Work
In this episode of 'a splice of Life Science marketing', hosts Matt Wilkinson and Jasmine Gruia-Gray explore how to create personas that go beyond static documents to become living, interactive tools. They discuss the evolution from demographic-focused templates to AI-powered personas that can help with role-playing, message testing, and understanding the jobs customers need done.
What Makes an Effective Persona
Matt Wilkinson
And welcome to a splice of Life Science marketing. I'm Matt Wilkinson and
Jasmine Gruia-Gray
and I'm Jasmine Gruia-Gray. Hi, Matt. How you doing today?
Matt Wilkinson
Good, good. It's a little bit steamy here in Northern Virginia,
Jasmine Gruia-Gray
quite warm here in the UK for once, which is nice. Today we're going to talk about persona. And that's probably quite an interesting way to start really is because you, if you create a persona of an English person, you might actually say that one of the things that they're always talking about is the weather, and you might say that they're always complaining about the rain. But before we delve too deeply into persona, it's probably worth starting with, well, what actually do we mean by and a persona is a really helpful way to represent a group of our customers. The way I like to view them is that within our ideal customer profile, we have a buying group, and within that buying group, we typically have people that are our primary buyer, and then we have users, and then we have those that are also involved in the buying decision. So that could be procurement, finance, health and safety, quality, whoever it is there's often quite a few other people that are involved in that buying group itself, and we can create across our different groups of ideal customer profiles, we can create persona. And those persona are these aggregated views of who they are. And so we can start off by thinking about trying to humanise our messaging around the roles that they play in the jobs that they do.
So for example, you might be looking at a lab, so a bench scientist in the lab, and so you know that first thing in the morning they come in, there's a certain number of tasks. Maybe they're checking their email, they're doing a few things, but then they're going into the lab, they're putting on their lab coats, they're putting on their safety specs, and there may be going on, there's certain tasks that they're getting on with. And so it's about understanding what their day looks like, what keeps them awake at night, and really, what gets them excited and what motivates them. And if we can understand that across a group of customers, we can speak more closely to their needs, wants and the jobs that they need to be done.
From Templates to Jobs to Be Done
Jasmine Gruia-Gray
Yeah, just sort of pulling on that thread some more on the jobs to be done. I think that's where persona development has really taken off in the last three, four years. And the analogy I like to use actually works really well from what you and I are wearing, I think in the before times when people thought about personas, it was very much a template. It was very black and white, and it was some cutesy name for the persona, what their age was, what their hobby was, all stuff that really, by and large, wasn't relevant for how they make decisions, what motivates them and what the jobs are to be done. Today, we're living in colour. We're living in a world where anything goes, where the templates are okay. I mean, I absolutely have my playbook of templates, but we think about the template of a persona from a much more human perspective, not only jobs to be done, but behaviours. What are the steps they take in making a decision? What are how do they think about what they really need, what they don't need, what they're fearful of, what makes them successful?
Matt Wilkinson
I absolutely agree. I think there's another thing that's really important about persona is that once we've created them, they can be a real pillar of coherence between sales, marketing, customer support, about around how are we going to ensure that our interactions with this group of people that we're tying up together as a persona within a single person, that we're going to make sure that we're trying to address their needs and wants and what they're trying to accomplish, and that we're able to really be able to provide better customer experience overall. I think that's the big thing. Because if we try to, if as human beings, we try to address every single person as an individual, we just can't cope with that. I mean, research shows that we can have maybe 100-150 friends before we sort of have to move out and sort of just sort of calling them acquaintances. There's a reason why communities kind of top out at the size they do. That's why organisational structures don't get too big before having bigger structures in place, because as humans, we just can't cope with that many connections. And so this is really a way to help us simplify that, but also keep it very human.
Jasmine Gruia-Gray
Yeah, I think there's a really important point you made at the beginning, which is that personas, and I'd add also that ideal customer profiles, the bricks and mortar, are really the unifying elements across a commercial organisation, and it's really critical that sales, marketing, customer experience service, FAS are all on the same page regarding the ideal customer profiles and the personas, so that you're all going to be targeting the same segments, the same potential buyers with a very similar message, and a message that you've tested, that you know is going to resonate with them.
Understanding the Complete Buying Group
Matt Wilkinson
Absolutely, I think that's why it's always so important to try and get those people around a table to agree on not just who are ideal customer profiles and what are they, but then who are personas and what do they look like. And I think that there can be huge benefit gained from going through that process, because you get that unifying kind of coherence around who's important to speak with, who do we speak with, who do we know, and also maybe getting insight about who don't we know, who do we need to spend more time getting to know you know, the number of times that you speak to clients, and the sales team, are doing really well. They're great at getting into conversations. They're great at putting in proposals. But then once it hits procurement, things go quiet. And so then maybe there's a conversation that we need to look at. What is it we need to do to make sure that the procurement like us as well. And so then, by creating persona around procurement, understand what is it that we're not doing that's not communicating with them. Maybe we need to get them in the room as part of our buying process. Maybe it's actually we just need some really simple sales enablement messaging that allows us to continue to have messaging that resonates with that persona.
So I think that's a really important thing to do, because we understand who's in our buying group, we can better understand what jobs, what's important to each of them, what jobs are they measured on? And so there's some really interesting ways about taking that forward, rather than just taking the very vacuous kind of approach of, well, they're going to be this old and be married with kids, or they're going to have these sort of demographic traits, but actually understand what are they measured on? And because each of the people in that buying group are measured in different ways, and I think that's a really important thing that we have to look at. So we have to try and create coherence within our buying group that helps make them that helps the buying group as a whole come to a decision.
How Many Personas Do You Need?
Jasmine Gruia-Gray
So in your experience, how do you think about this? Is there an ideal number of personas that a company should have, or a product group should have.
Matt Wilkinson
So that's a really nuanced question. I think, in the days before AI, or at least before Gen AI, I would have said that trying to understand, in detail, who your key people you need to influence, and really digging into making sure that we're creating, we've got messaging for maybe, typically, there are five roles with any buying group, we sort of listed them out earlier, sort of the initiator, the user, the initiator is kind of the problem that recognises the problem. You have a champion, who's there that's also part of that similar group, but is there and sort of willing to take the drive to really solve the problem. You have end users, you have procurement, and you have finance, so you sort of have these interested parties in completing sale. And of course, those roles will change. They will be different depending on who you're selling to what the organisations are like. But I would have said, start off with any particular type of sale, with having about five persona within a segment, within an ICP group, however you're creating that segmentation.
The problem I've always found with persona is that very often what will happen is that agencies will go away. You'll have a fantastic meeting where you'll get everybody around the room, great workshop. Everybody's really enthused, and they're agreeing on things, getting these, the notes that come back sort of half formed. People then go away do some desk research, spend hours kind of trying to fill in the gaps. And then once you've created these, they get turned into beautiful documents that then just rust away on a hard drive somewhere. They get used for the initial piece of work, and then people forget them.
Using AI for Deep Persona Research
Matt Wilkinson
And so a deep research came out. I realised that rather than doing the manual work yourself of going through LinkedIn profiles job descriptions, you could actually create a way of using the deep research tools themselves to go off and analyse and think about a group of profiles that are on LinkedIn that fit your persona, looking at their job descriptions and then really trying to enrich what the conversations that you'd already had with genuine data that could really help to help you better understand who you're speaking with. And so that's something that I've been using now for, I guess, the best part of a year.
And then on top of that, rather than just having these sort of persona documents that are having to analyse and read through and go right, what was their key questions that they asked as part of their buying journey, what matters most to them? We could start creating custom GPT or AI assistants where we're actually uploading those into as a knowledge base within the AI itself, and then using that to not only be a queryable kind of persona that we could actually interact with and ask questions of rather than having to us ask the questions of the document. We can just go into the AI and ask the chat bot ourselves, but we can take that a step further and then actually interact with them and use that as kind of as a an in silico version of our persona themselves to do everything from brainstorming, testing messaging, creating messaging, the whole gamut of the sorts of things where you can almost have a virtual customer at your desk all the time. And I think for me, that's been one of the real eye opening ways that I've certainly changed my workflows, and I think that we can really look at making these personas so much more useful and helpful and shareable.
Creating Unified Commercial Teams
Jasmine Gruia-Gray
I can't say enough about this. I think this point is so important, and links back to what we were talking about earlier, about creating unification across a commercial team, because once you can agree on the jobs to be done and on, as Adele Revella calls it the five rings of buying insight, priority initiatives, success factors, perceived barriers, buyer's journey and decision criteria. And then you operationalise that to an extent, as you say, in silico with a custom GPT or a gem, or whatever your favourite LLM is, you can now use that to create sales scripts, to create sales presentations, to really evaluate your website, to evaluate your marketing tools and your sales tools to helping your customer experience team, because now you're all working on this using the same platform to think through what that outbound messaging is going to be like, and what that experience is going to be like from a proxy persona.
Matt Wilkinson
Yeah, absolutely. And I think that especially when you've got the ability to share these now, having everybody using the same data source or the same AI and silico version allows you to have consistency across the silos. And I think that's then really important. Now, of course, one of the criticisms has always been a persona has always been that there are one and done. I don't think we're quite there yet, but you can imagine a future where you build these persona into an AI that sits within your website, you can imagine that then you're learning not just on what does what we're telling it do within a website, but what actions do they actually take when we can recognise a specific persona? And then, if we're at that point, can we use data collection to then enrich them? I'm sure that there's, there are ways to do that. I, clever ways. I don't think there's a product out there that does that just yet. And I'm sure that HubSpot and Salesforce are probably working on this as I speak. But I'm sure there's ways to do that.
Role-Playing and Voice Coaching with AI Personas
Matt Wilkinson
I think there's also another thing I've seen is that there are now these speech coaching tools that companies like use it usually, and a few others, where you are able to not only understand to look at your customers and understand them better, but you can build those persona into a voice, into a conversational AI, and then you can set up scenarios. So let's just imagine that we've built our AI, we built our persona for a lab scientist. If I'm a salesperson, I'm going into a call with that lab scientist. Maybe I take that same person and build it into a conversational AI, and within that I can then actually practice a specific scenario. Let's just say that I'm going in and meeting them for the first time and I'm trying to introduce a new product range to them. I can actually practice that conversation. I can practice that cold call. And so there's a whole wide range of ways that, once we've got this, these data sets, that we can start using them to not just create better and more, but also to be better and more with them, to actually be able to practice with them. And I think that's just one of the really exciting things that I'm seeing.
There's also ways to look so get buyer insights around their personalities. And so you could imagine that there are different DISC profiles or Big Five personality traits. And you could look at those and go, okay, so if we've got these different types of personality, because we know that people aren't all the same, even if they're doing the same job. Well, we can then look at what would it be like if that person is more of a D rather than an I in their DISC profile, and be able to practice those same conversations with different personality traits. And I think that then just takes things to a whole different level, where all of a sudden we can be more prepared for the conversations that we might want to have.
Jasmine Gruia-Gray
Yeah, the word that comes to mind when you're describing this use case is informed, so it will help. It's the classic role playing that we used to do at sales meetings, right where somebody in marketing would play the role of a customer, and somebody in sales would have to practice their pitch, but the scenario you're describing with an AI helps you actually build something that's much more informed, much more real world and more scalable, so that it doesn't just live in the sales role playing, it can live in the customer experience role playing and how they deal with different scenarios and being on the front line as they often are.
The Importance of Real Customer Interviews
Jasmine Gruia-Gray
I think the other point I want to pull on is that just because we can train LLMs to be a very good proxy for customers and for personas, that doesn't negate the need to actually do the hard work and do some of the interviewing. And I really want to encourage our marketing colleagues and listeners to get out there and get that experience and roll up your sleeves and really understand face to face what that buyer's mindset is, what that buyers context is and how your company and products and services can be useful to that buyer firsthand.
Matt Wilkinson
There's nothing better than that firsthand experience of what it's like to walk in the buyer shoes for even if it's just a few metres or a few miles. But I think there's something else we can do as part of that, if we're given permission, if we can record those interviews, or we can at least take notes and add to them, we can add those verbatim quotes, those transcripts, to the data set as well that we're going to add in. So rather than just having a very polished persona, the AI doesn't really care how much data we give it. So we can then go through and actually add transcripts after transcript to these that will then actually only enrich the data. So we can constantly update these by simply by adding extra data to them, whether you're updating the same file or just adding extra files to say, hey, here's a couple of transcripts from interviews with this persona. You can do all of that and really start to build out a wealth of real customer voice, and have that shining through in actually, the way that they might speak, in the way that they might want to be spoken with. And so I think that's a really powerful thing.
Another reason to go out and speak with customers, not just because it's great for each of us individually to get that sense of that humanity, but because we can add that in, and we can then share those insights across the entire organisation through these in silico versions of our customers.
Jasmine Gruia-Gray
Right? And it's not only your customers that you can interview. You can also interview folks that didn't go with your solution, that chose your competitors solution, or chose some internal solution.
Matt Wilkinson
Yeah, absolutely. I mean, I use the word customer, but I think it's target, or potential customer, prospect, whatever we want to call them. I think that there's absolutely that it's really important that we do get that sense of, well, who's buying from us, who isn't, and what are the differences? Is it just because of brand, or is there something inherently different about the way that we appeal to certain groups of people and not others? As you know, when we're selling things, some decisions are made purely based on the procurement level, where you have a single supplier and or you have a group of suppliers, and if you're not part of that supplier list, you're not in the game. And so then it's a case of, well, how do we get in the game? But being able to just even understand that, I think, can be an incredibly powerful part of this sort of conversation.
Common Persona Mistakes
Jasmine Gruia-Gray
Right? So just before we wrap up as a take home message, what do you think are the two biggest mistakes that marketers make when they create buyer personas?
Matt Wilkinson
So I think there's so many mistakes we're about creating and then, but I think the biggest mistake is that often personas get created and then they used and forgotten about. So I think the first one is that they get forgotten about that, I've seen some beautiful persona being created, and then people aren't even using them when you're having to remind people to go back to the persona, because the information is in them, so they're not using them as a source of truth. So that's the very first thing to remember, is to use them. And I think that by turning them into an AI representation of those personas, actually makes them so much more useful, you just get far greater ROI for those efforts. Anyway. So that's number one.
And I think the other is, by being a bit too shallow with the questions that you're asking, I think that the five rings of buying insight so powerful understanding what jobs to be done, any of that insight that we can really get included? Yeah, there's things that are important about what channels they want to communicate through, but the more important thing is, is actually, how do we move them emotionally? How do we actually meet what their needs, the jobs that they need to be done? How do we actually help them achieve what they're looking to achieve? And if we can answer those questions within our persona, all of our messaging will then be just so much stronger, more focused, and will resonate so much better. And that's really, I think the big thing that I'd focus on is make it so that's usable. And it's not just pretty fluffy pictures. It's actually something that really can be used by marketing, by sales, by service, and anybody else that's interacting with our customers and prospects.
Starting with Personas in Campaign Development
Jasmine Gruia-Gray
To sort of add a finer point on everything you've said, and I agree with is I sort of appeal to product marketers and marketing managers when you're creating your next campaign, start with the persona. Who is it that this campaign is appealing to, what are their jobs to be done? What are their motivators? What is their buying context? And then put the campaign together, and then after the campaign is executed, and you look at the data, who has responded to your campaign, not only from a persona perspective, but also for an ICP perspective. Are there any surprises there? Are there any outliers from the trends that help you to then loop back to your ICP and to your personas and tweak them and lift them up even further?
Matt Wilkinson
Yeah, or even find that you've got a segment that you didn't realise that you were actually, that you were able to serve, that needs to be treated differently. Maybe you've hit an outlier that actually should be a separate segment, and we can do even more with that with those outliers. So I think you're absolutely right. I think being able to treat everything a little bit like a an experiment and a hypothesis. And knowing that very much like in science and in chemistry, when I went to do my first year of chemistry, basically they told us everything that we'd learned up to that point was wrong and that the models we'd use were all too basic. And as you go through doing organic chemistry and inorganic chemistry and physical chemistry, all you're doing is applying a model to serve a certain to solve a specific problem. If you try to use the most complicated model, the kept scientific model here, I'm not talking about large language models, but the most complicated scientific model to solve a simple problem, you'll be just wasting too much time, but you'll get the same result 90% of the time, 99% of the time. So what you're really looking for is finding the appropriate model.
And I think this is really the trick here with persona, is that they are an appropriate model for a whole range of communications and helping us get communicate better with a group of our customers or prospects. But they're not perfect, and of course, we do need to layer on top of this personality and individuals and the individual context of every customer. If we were taking to this to Account Based Marketing and a key account, we might create individual personas within a market of one company. So I think we can, and with AI, the way it is now we absolutely have the ability to do that. We can scale these things far quicker, and then we can use these in a way that actually makes that make sense. Whereas if we were doing this on a basis where we're looking at hundreds of persona potentially, but you're having to do the analysis, and kind of, you or I are having to try and understand the difference between somebody in the same job role in two different companies. We wouldn't be able to necessarily do that. I think the subtlety of that's different. But within the AI, maybe actually within some of our most important accounts, it's worth us doing that. It might not be, it's a hypothesis that we can test. Does it make sense to do that? But if we treat everything like an experiment, where we start off with a hypothesis that this is a good this is going to be good enough to make a difference, and for us to be even just 1% better than if we don't use those, that's where I think it's just so important.
Jasmine Gruia-Gray
Yeah, I think that's fantastic, and I think that's a great message to end on. This has been tons of fun. Thank you for sharing and helping us all learn, Matt,
Matt Wilkinson
and thank you, Jasmine, it's been great fun as always.
Jasmine Gruia-Gray
And thank all of you for joining us on a splice of Life Science marketing. Bye, for now.
Matt Wilkinson
Bye.
