Data therapists are translators that sit between business and technical data leaders. What happens in a data therapy session? Who runs it? When do you know you need one?
Join hundreds of practitioners and leaders like you with episode insights straight in your inbox.
Checkout our brands or sponsors page to see if you are a match. We publish conversations with industry leaders to help data practitioners maximise the impact of their work.
Data therapists are translators that sit between business and technical data leaders. They are great facilitators, they understand tech but they also speak business. What happens in a data therapy session? Who runs it? When do you know you need one? Today I learn from a real expert on this topic, Ashley Faith.
Ashley is a professional data scientist with over 15 years of experience focusing on knowledge graphs, taxonomies, and machine learning. She's also the host of the Knowledge Graph, Data Science YouTube channel under the channel name Ashley Faith. She has worked with GM, NASA, NATO, Down Jones, IEEE, EPSCO, and the Department of Defense among many others.
More from Ashleigh
Ash has a fantastic YouTube channel check it out here:
Events coming up - Data Management Marathon 5.0
The Data Management Marathon 5.0 is coming up soon October 12-13,2022. This is a one-of-a-kind virtual event for every data lover, born out of a love for data management, knowledge sharing and storytelling. Organized by Thinklinkers in collaboration with the one and only Scott Taylor, this event features high-level speakers, influencers and an enthusiastic community of data professionals.
Check out the event agenda and use our special promo code DISCOVER20 to get a 20% off the pro pass:
Your ideas help us create useful and relevant content. Send a private message or rate the show on Apple Podcast or Spotify!
**Loris Marini:** We are here to talk about building bridges between business and technical data leaders and a great way to do that is to run a data therapy session. Data therapy is something that you might have heard a couple times. I certainly did, but I never actually had a time to dive into the topic and really understand how does that work?
What is what is the structure? Who runs it? When do you know you need one? And so today I'm sitting with a real expert on the topic, Ashley Faith. Ashley is a professional data scientist with over 15 years of experience focusing on knowledge graphs, taxonomies, and machine learning. She's also the host of the Knowledge Graph, Data Science YouTube channel under the channel name Ashley Faith.
Which I absolutely recommend subscribing to. I did it recently and oh my gosh, how much content really useful. But as usual, links are gonna be in the show notes, but just a bit of context there. PhD, Ashley's PhD and continued research efforts focus on advanced semantics and bridging the gap between users and inform.
Which is what really gets me out of bed in the morning. She has worked with gm, nasa, nato, down Jones, IEEE, EPSCO, and the Department of Defense among many others. So it is an absolute pleasure and to have you on the show, Ash. Thank you for taking the time and welcome to the podcast.
**Ashleigh Faith:** Thank you for inviting me. This is fun,
**Loris Marini:** Yes. So let's start with the basics. What is a data therapy session?
**Ashleigh Faith:** Yeah, I mean it's funny to call it a DA data therapy session, but that's how it makes you feel, right? You have a facilitator. And then you have the parties that you are facilitating. Usually they're the business or product side, and then [00:05:00] you have the technology and engineering side.
They don't often know how to talk to each other very well. They want to desperately, but they just don't know how to communicate well with each other. And that's what the data therapy session really is about, because you have a problem, you have a project, and you need these two groups to really understand and be in lockstep with each other.
but because they can't communicate, they oftentimes talk across purposes. They get upset over things. They highlight certain things that maybe are not appropriate to the larger audience. And a DA data therapy session is to work through those issues and also to collaborate. It's not always about issues.
Sometimes it's about, Wow, did you know we had this cool data? Wow, yeah. We can totally use that to make, this kind of revenue generation on this. It's really eureka moments that you're looking for in these as well. So you have these when you need them. I know a lot of companies now are doing them on a regular basis, like your daily standup sort of thing, and they're becoming very effective.
**Loris Marini:** Right. Cool. That's good news because we definitely need to build those bridges and, but I wonder as a team, when do you know that you can need a data therapy session?
**Ashleigh Faith:** Yeah
**Loris Marini:** indicators is.
**Ashleigh Faith:** yeah, leading indicators. So normally I always like to do a data therapy session before any kickoff of big project. So if you're using, any kind of agile methodology, this would be during like the planning stage. where you're really trying to get acceptance criteria on the business side and on the technical side.
So the business side, right? You're talking from the user or the customer's perspective. And oftentimes the engineering and technical folks don't know who they are. They have no idea about user persona and that sort of thing, and vice versa, right? Like the business side doesn't often understand when you're talking about data architecture.
It's a process. It's usually a pipeline of some sort. So they have to understand that just because they need it in three months doesn't mean they're gonna get it in three months, and you need to help them understand why, right? You all have to be on the same page as to the why these decisions are being made and why certain things can or cannot be done.
So usually it's like at least do it as a kickoff onto any project.
**Loris Marini:** So what's the difference between a data therapy session and a brainstorming session or an ideation meeting?
**Ashleigh Faith:** Yeah. The it's the intent, right? So when you get into data therapy, you wanna prep everybody ahead of time, that's what you're going to do. Because if you don't do that, you're gonna get the idea session, which is everybody trying to get their idea. Clout and everyone voting that their idea is better maybe than other ideas.
When you talk about it from a data therapy perspective, you're saying, Look, we are not coming into this and talking about the how. We're not talking about the how. We're talking about the why and the what. That is why this is so important before you get to ,
**Loris Marini:** Okay.
**Ashleigh Faith:** how stage. Because, and speaking as a technologist myself, oftentimes the business wants to tell you how it should be done.
And we all know that doesn't work very well, . So having this allows everyone to that moment where they can say I think it should be done this way. And they go along with that for a few minutes, let them get that out. And you say, Great. That's not why we're here. . So they get it out and you all start to get onto the same page as to what are we actually trying to achieve?
Because on the business side, they're not trying to achieve creating a new knowledge graph that's a. , right? What they're trying to do is understand that they are retrieving the correct content in recommendations, and here are the ways they're going to measure that. That's great. If you take how that's gonna be measured, the technology side now has to say, Okay, cool.
So do we have what we need? Do we have the data? Do we have it set up to do those things? So it centers everyone about what are we actually trying to achieve and everyone stay in their own lanes when we do it
**Loris Marini:** It's funny cuz when I think about therapy as someone that has started a therapy journey with my own therapist which is opening lots of interesting new ways of thinking and, really sometimes putting me in front of realities of my personality that I'm not particularly proud of.
I think about that, I think about that safe space where you get to drop your shield and openly discuss what it is that is traveling you. Why you can't sleep, why you find yourself awake at 3:00 AM for no reason, right? Apparently no reason. Hopefully all the listeners to this podcaster have a sound sleep time, [00:10:00] but I'm sure there's some of them under pressure, particularly in industries that are under the pump with all the post covid, the shenanigans we are witnessing.
Money is tight, money is more expensive, there's more pressure from the business. It's understandable, right? If you lose a sleep a couple times during the week, what you just described sounds to me like a data strategy session, what a consultant will call data strategy. I get the business, I get the technical folks, we talk about the what and the why,
**Ashleigh Faith:** Yeah.
**Loris Marini:** and we don't necessarily cover the how.
The how is come to later once we decide. Do you think there's a difference between, this did a strategy session and a did a therapy session.
**Ashleigh Faith:** Yeah, there is. They can actually overlap quite a bit, but the way that I always conduct these is in these sessions I take on the persona of the therapist. So you allow that discussion to happen and. While people are having that discussion, you tease things out When you start to, the thing that you were just talking about, the things that keep you up at night, the things that you don't necessarily feel.
Saying in a different space, this is what you wanna really highlight. So having a person, or, maybe you tag team it with someone is I think the critical piece here.
Let's take a machine learning example. That, that I recently went through. We're trying to classify the document types for this corpus of information. So it's just we hear from the technology said, Oh, yeah. It's just a classification model that's easy to do. And I say, Hold on. So as the therapist, Hold on a minute. Do you know, have you seen the data that you're gonna be training on?
Oh, no, but I'm assuming half
**Loris Marini:** oh, this
**Ashleigh Faith:** We're assuming something. I'm assuming that the business side is the SMEs. They know what they're talking about. . So what happens is I say, Okay so business side, how confident are you in your own tagging? And they're like, Oh, it's actually really historical and multiple people have done it over the years, so we don't really know you know how trustworthy it is for machine learning. We don't really understand how classification works.
Aha. So now we understand that the technology side was saying, Oh, it's an easy classification, right? They dismissed it as something easy. I totally understand this. On the business side, they don't have any trust in that because they don't know what a classification model is, so they can't really tell you how confident they are with that solution or with the data they're going to provide.
Long story short, in this, we decided to look at the data together and we discovered that the training set. was all over the place that the SSEs could not even determine the difference between different classes. There were over 95 features they were trying to train on. And so learning this early helped us understand the true problem that we had to solve, which was, we don't even know what categories of content we.
So let's start there. Instead of going into this machine learning model and reiterating and everyone getting frustrated and it not working out in the end, and it's because we had this session where people were like wait, what? What does this, You're saying that you're highly confident in this, There's a confidence score.
What does that mean? And it's that it's half information literacy, half data therapy.
**Loris Marini:** As you were telling. Of that particular problem, a book came to mind is maybe you read it, is The Coaching Habit by Michael Stanier.
I love that book because it really goes through systematic seven basic questions to get to understand the real problem and get to meaningful conversations.
And I think the underlying theme is really to create space, right? Like we live in. Constantly on the pressure. Time is of the essence. Everything has to happen so quickly, and we are, We used to think that our value add is associated to our ability to know what to do. which as SMEs obviously you would expect, a subject matter expert should know their domain.
But solutions to business problems rarely involve one person alone, like a heroic effort solving the problem just by themselves. It requires collaboration and there are many interdependencies and we need to get to understand the landscape before we start training and doing
**Ashleigh Faith:** Yeah, absolutely. There's a story I have around that as well as to why this is helpful. I think oftentimes people are scared if they're, especially if they're in senior roles, in larger groups to say, Wait a minute, I don't know. What does that mean? Because to your [00:15:00] point, they feel like they should know the answers because they are, higher leadership.
So these data therapy sessions, because they are smaller, right? And you can get the folks in the room that need to actually talk to each other. I was in a meeting where we were going over a 40 million project. Okay, so it pie stakes and I was the data architect that was in the room for this, and this was many years ago.
And we were talking about how we were going to take. the data we had and start to map it into more standardized schema so that, we our whole ETL process was going to run smoother. And I said you don't necessarily need to pick everything up and remap everything. You can actually use graph to, to connect to the data sources you already have.
And to do that mapping that way. And I said it's using linked data properties. So again, this was many years ago before a lot of people really understood what Link data was. So the senior architect in the room assumed he knew what that meant and he's Oh yeah, that's just an implementation detail.
I'm like, Wait a minute. That doesn't sound like an implementation detail. So we had a session right after that big meeting, and what I discovered was he thought Link data meant just hyperlinks and not. Open Link data it was closed to, to the company, but link data principles. So having, unique IDs and going with graph, and because he made that assumption, it would have set us back two entire years worth of work because he hadn't accounted for what we were actually talking about.
So having that session with him after made him feel like he wasn. On display. His knowledge was not incomplete and it actually gained a trust between the three of us that We were talking in the data therapy session because he felt that I wanted to understand his perspective.
**Loris Marini:** Yeah.
**Ashleigh Faith:** the other folks that were gonna be developing this, needed to understand that the two major stakeholders, myself and this architect were on the same page, so we added more confidence to the project.
And it's a little bit of that softer skill stuff, but that's what will really make or break a project is if people are questioning, Wait a minute, do these two know what they're talking about? You don't want any large project having that issue.
**Loris Marini:** I think, and this is critical, I think the skillset for running a successful data therapy session, you've gotta have that emotional, intelligence there to understand what it feels like to be the person you're talking to and what might be. In their heads, of course, we don't have a crystal ball, but asking a question is a very effective way to get an answer.
And if the answer doesn't seem satisfying, you can always ask another question. but there's a way to do it. I've, I remember being in sessions where, if you just transcribe the conversation and read the exact word, sequence of words, those seemed pretty innocuous questions, from someone that wanted to generally know more, but the body language and the way they were framed were saying the opposite.
Were questioning or. Either trying to protect their own ignorance or maybe with hand waving approach getting to the answer, jumping all those. assumptions, those moments where you take a deep breath and hang on a minute. What am I assuming here?
Are those assumptions agreed upon by the people in the room? Are we on the same page? It it feels weird because you have to slow down almost, right? You have to.
But does that come natural to you or you've gone through a process of adjusting and understanding your pace? The type of body language that tends work more.
Is this something that you learn? Because I feel like any data team or every data professional should know how to run a data therapy session. Cause if you don't run it, you're gonna be part of it.
**Ashleigh Faith:** Yeah. Yeah, absolutely. I will say that I do have a natural affinity to be empathetic. I think that is a big part of it. That doesn't necessarily mean that you can't learn about this. Cuz I've had a lot of learnings as well. I used to be the person that would be like, Yeah, okay, we can do it.
This is, I was so excited and so yeah, let's do it. Let's do it. I am always passionate about these things. I've had to reign myself in on a lot of this. but one strategy that I, two strategies, I would say that helped me immensely. One is I tend to talk fast, and it's because I'm passionate. I wanna go, I wanna move. But I need to take a breath. You said that earlier. Just after you just take a breath.
I'm also one of those people that likes to fill in the silence. If you're one of those [00:20:00] people, make sure you take a breath, because if you don't, you will be the only one talking. I have had to learn that skill. The other thing that I know has been incredibly helpful, and this is like I would say one of my biggest pieces of advice.
I worked with a different data engineer. That was constantly assuming things and he had a reputation for that.
**Loris Marini:** Uh huh.
**Ashleigh Faith:** And what I asked him to do is, his name was Ollie. And I said, Okay, Ollie, when you think you know what I'm talking about or with anybody, right? Cause he was having this issue with a lot of different people, repeat back, right?
So somebody says something to you and you say, Okay, let me repeat that back to you and where I get it wrong, you correct. That it's so simple, but it helps you understand because then people will correct where, because people hear what they wanna hear, right? People misinterpret what you're saying all the time, and this is incredibly helpful, especially when you're talking to your technology stakeholders because they might be assuming things about whatever you're talking about.
And on the business side, they might not understand the words you're using, even the word database or customer, they have so many different meanings and a business person might not know that.
And so having this, Okay, let me repeat that back and see where I get it right and where I get it wrong. It, I would say that's the biggest thing that has changed the way these.
**Loris Marini:** I love it. Two very simple, two very simple and applicable tips. One, slow down, take a breath, breathe. Inhale oxygen, wait for the brain waves to sink and then continue. And two repeat. With your own words, try to summarize the concept not in a way that sort of implies that you got it, but in a way that generally asks, Is that what you meant?
Did I understand correctly? Not even what you meant, because what you meant implies something about the other person. The intention,
**Ashleigh Faith:** Mm-hmm.
**Loris Marini:** but just make it about yourself. You are a receiver, There's a transmitter. Did I go get it? ? yeah, we need to know more of that.
**Ashleigh Faith:** Yeah I just, and here's the thing the other anecdote to that is when I told Ollie about this, he actually came back to me. He said, you were the first one that actually took the time. to communicate with me on this. And I think that, I love that idea. And it changed his entire outlook on how to work with people.
And I'll be honest, he was in some hot water at work because of the actions he was doing. And he wasn't really aware of it because nobody had taken
**Loris Marini:** Oh, that's a
**Ashleigh Faith:** talk to him. and work with him on this. And after that, like this was we had one of those things where when you do performance reviews, your colleagues get to give you feedback.
It was during that time that, that we had this conversation he just loved it. He, and he gave me that feedback to, and this was a pretty hard guy, right? Like he was not the type to come back and say, Too many nice things, but for him to come back and say, Wow, thank you for that. That's a pretty powerful statement coming from someone that you're trying to help.
**Loris Marini:** This is coaching for development. It's about help people express their ideas. And their work in the best way possible. Trying to get to be more mindful of your own limitations. Cause everybody has them. it's just a matter of seeing them sometimes, as you said, there's no time.
so part of perhaps the data therapist mandate is also to create that space intentionally so that slow things down and go Hey, there's no rush. I know Slack is pinging,
**Ashleigh Faith:** Yep.
**Loris Marini:** but can we put our MacBook Pros away for a moment I was rethinking about what happened to me that day in that startup, in that room with those people. Everybody was hiding behind the keyboards with them MacBook Airs, and Slack was pinging, and you've got, not everybody was on mute, so you got this and you're like, you're trying to think, you're trying to get to the bottom of a problem.
And Slack pinging continuously on the background is not the right environment, right? So those are clue as well for who runs it, perhaps a session. Be mindful of the set and setting. It almost sounds like a psychedelic trait, but it is about set and setting, It's
**Ashleigh Faith:** No, actually that. That's a big part of it too is it's almost like you go into one of those sensory eliminating pods sort of thing
**Loris Marini:** Mm-hmm.
**Ashleigh Faith:** because you really and it doesn't have to take a lot of time. These conversations, they can be 10, 15 minute conversations as needed, so they don't have to be these long, drawn out sessions.
I've had those too, believe me, I've had those where there are giant whiteboards. There's people getting heated with each other. I've been in those two But these do not need to and [00:25:00] actually I would say have these early and often is a good rule of thumb because then you eliminate you these buildups, right?
The whole put it in a bottle. Okay? We just need to, I'm irritated with this issue. I don't know where this is happening. This thing is happening and we need to get to the bottom of it, but we're on a deadline. Just shove it in the bottle. Shove it in the bottle.
You need to have a moment where somebody that is in your project is maybe the data therapist lead in, in this situation where people can come to them and just say Look, I'm having these issues.
I'm actually really excited about this thing, but no one seems to understand it. I don't know why I can't get through. I don't want, I wanna make sure everyone listening knows this is not just when you have issues, it's when people wanna be heard and. I am no surprise, one of the people that is very open to just talking and being an extrovert in meetings.
But there's a lot of really smart, really talented people that don't feel that way. And, maybe there's some more junior level people in these larger meetings or these larger discussions that don't feel empowered. To talk. Even if you are empowering them, they're like, Oh, I can't say this to the director or so and
The a person to be able to go and talk to and then facilitate some easier conversations.
There was a very junior right out of school machine learning engineer that I was working with, that he had this amazing idea. To generate relationships in a knowledge graph. But he didn't want to bring it up in this meeting because the VP of technology was in it.
And he's I don't know if I'm allowed to talk to this person. And so normally your stakeholders are very open to talking. They just have very limited. . So I set up a meeting between the two. A data therapy session and prep is key to all of this, is making sure everyone knows exactly what you're gonna be talking about, who's going to be involved, and what are we expecting to come out with, right?
So I know these sound very standard of course, but oftentimes when you're having these little short conversations, you forget about it.
**Loris Marini:** So go through that at least again. So three important things. Who is involved in the meeting? What are exactly, what are they gonna talk about? And the third point was,
**Ashleigh Faith:** what you want to get at, What do you wanna
**Loris Marini:** The outcome, the success, yeah.
**Ashleigh Faith:** we had a short email exchange where I said, Hey, so and so has a really great idea. They're just a little worried because they're new. You're pretty high up in the company. I know you're open to talking to people let's just have a small meeting if you have, 15 minutes and.
I knew it was gonna take 15 minute conversation because I had talked to the original, engineer that wanted to have this conversation. So the data therapist, you can see this as maybe taking some time. So I would say that
if the person you are thinking of right now that could be a data therapist oftentimes they are the people that have to understand the translation, like the translator between.
Anything. Usually these are maybe like the taxonomists or the people that do modeling or, even like the marketers are really good at understanding, okay, how do I synthesize this into normal people
**Loris Marini:** Yeah.
**Ashleigh Faith:** So these people exist at your companies already. It's gonna take them some time to get up to speed on how to do this, but once they are.
You, they start to identify when these data therapy sessions might be helpful and coaching your senior leadership on it too. I've had senior leaders come in and say, Ashley, we really just need to get these two really strongly opinionated engineers to see eye to eye on their solutions because they're solutions we think are complimentary, but they're talking across purposes.
I'm like, Okay, I'm on. And in that situation there was some prep work where I just sat down with them and said, Okay, what do you think the problem actually is? , where do you think the problem
**Loris Marini:** point. Yeah.
**Ashleigh Faith:** Yeah. And that's the thing, be direct, going in and being like, Okay, we're gonna just talk a little bit.
You have to be very direct, very
**Loris Marini:** Yeah. Yeah. Especially with engineers, they lose patience very quickly. Yeah.
**Ashleigh Faith:** Yes. I've worked with engineers, been an engineer my entire life. And so I, I think like that. But I also know that to get your job done, you need to have that stakeholder buy-in and they are not normally technologists. And you need to know and understand your end user, who are also usually not engineers or data
**Loris Marini:** You touched on so many points, actually. So one, one big one. I would perhaps summarizing [00:30:00] correct me if I understood correctly
**Ashleigh Faith:** we go. Yeah.
**Loris Marini:** practicing,
**Ashleigh Faith:** it in practice.
**Loris Marini:** walking the talk the concept of psychological safety. So creating an environment where people feel safe to speak up and give, Put their ideas out into the world at the risk of sounding stupid or perhaps outta context or not prepared or inappropriate.
Sometimes all of those things can happen at the same time, but in a way you shouldn't care because good ideas often come from. Crazy last second bulbs that go up that light up. So you need to have that how do we create a psychologically safe environment as something that leadership coaches have talked about for a very long time is something that we need to, it's a skill that we need to absorb. And I think it's pos everybody can do it, but at the same time,
I have many memories of interactions with very technical people that speak like they tweet. Did ever happen to you?
You go in a room and you're like, So what's the problem? And they give you that one liner that sit, like there that you don't get more than that. So you go, Okay, so tell me more. Tell me more about that particular one liner that you just dumped at me, cuz I want to know a bit deeper. And you get another one liner. So it's, it literally feels like you're on Twitter. It can be
**Ashleigh Faith:** know, the, a, A good way to think about that is if anyone has children in their life,
But why? But why after everything you say, like, Why is this guy.
**Loris Marini:** You're
**Ashleigh Faith:** well, but why,
**Loris Marini:** you're saying be that kid
**Ashleigh Faith:** the same thing exactly that you're talking about. And I often say, Okay, let's tease that out a little bit. You have to start and I would say that this is the skillset to, to brush up on is how to ask good questions
Listen to something and understand what can I tease out of that is not service level? What is implicit in that is not written out and actually writing it out. On a whiteboard, I've done this, right?
Write it out on a whiteboard, underline all of the jargon or all of the things and say, Okay, how would you define this? How would you define this? What is this? Can you gimme more detail on this? Who owns This is another good one, right? When you're asking these questions, and if you write it out, and it doesn't even have to be on a whiteboard, it can just be through email or something.
They're going to get irritated. , right? But you have to keep the, their eye on the ball, which is, I understand that this makes sense to you, but you have to understand that there are new people on this team. There are people that have not been at the company as long as you have. There are people that are trying to do an effective job, and they are struggling to understand.
So let's all come to the table, right? So you need to have that facilitation. Any kind of facilitation skills is going to be a huge benefit to this. because they're not gonna wanna take that time and they don't have to take too much time. You have to be respectful of their
**Loris Marini:** Yeah. Yeah. But sometimes it's about, the net month or seconds you spend in room, It's, is the fact that you're there and someone is asking those questions in the first place. That sounds irritating. Even if it's an email, one question, three seconds that can Yeah can be a blocker or can be perceived as a lack of trust.
Or maybe you're questioning the ability or the knowledge of that person. Maybe you don't trust them. Because they already gave you that one tweet answer. You can walk away with that. In principle, that answer contains all the ultimate truth about that particular technical problem. But so what is it about you?
That is, is it because you don't trust me? So I feel. That inside, that word that you mentioned, facilitation. There is a lot. There is a lot in, it's a box, it's a Pandora box, and inside there is a whole bunch of skills that, to be honest, I don't know how to enumerate. There's emotional intelligence, there's empathy there is being able to listen and slow things down, but also be curious to not be afraid yourself of asking.
The seventh time, let's dive a little deeper. Let, give something, I'm not sure I understand. And so slowly, sometimes it's easy. Sometimes you ask the question and the person goes boom. And gets at the core of the problem excruciating details.
Sometimes too much detail, but it gives you the whole story, , right? And you don't have to, you're like, Okay, done. I got it. . Great. Let's go for coffee. That happens, but some other time. And that's easy because you got your answer, You've got your context. [00:35:00] That tribal knowledge, or really in this case, implicit knowledge is be made explicit by the person directly.
But yeah, a lot of times is about, pulling that rope and try to get, what is it? What is it? Let me see. Is it a small fish? Is it a whale? Is it a shark? What is it?
**Ashleigh Faith:** and then there's a good approach that I've learned for that if you're starting out, and that is ask the questions that have the most meaning to that individual, which are usually what's the highest risk, what's the highest cost if we get it wrong? , because that's usually where they're going to have lots of things to tell you,
So tell me about that one tweet. What is the highest risk in this, right? Because that's the thing they're going to have the most opinion on, and you're gonna tease more out from that. Are you gonna get the full answer right away on just those two? No, but in a way, If you are being the data therapist you yourself, have to start gaining that trust that you're not asking stupid and pointless and time wasting questions.
You're actually doing it for a purpose I can tell you that every single job I've ever had, everyone has said that this is the part that they enjoy working with me on the most. I am super smart at all the technology stuff that I do, but one thing I'm very good at, Is helping people get on the same page and it sounds like the most of course that's easy to do, or, Oh it'll come out in the wash.
It doesn't and honestly, I have seen where it can go off the rails very quickly and, What I have heard from my colleagues across all of the jobs I've had is they feel comfortable coming and talking to me. They don't feel like they're going to be treated as stupid or, not as smart at whatever they're doing because I don't judge.
I think that's where you have to, you just leave your judgment at the door and say, You know what? And I say this all the time, I don't know what I don't know. I could be an, I actually hate saying the word I'm an expert in anything because I don't think that's true of. Everybody can learn something new for whatever they're doing.
And if you go into it knowing that you're not the smartest person in the room now, if you think you are the smartest person in the room, there's a whole nother podcast on that
**Loris Marini:** How not to think that you're the smartest person in the world
**Ashleigh Faith:** and why you're probably not. And if you really, truly are, you need to find another room, right?
**Loris Marini:** Yeah ,
**Ashleigh Faith:** right? Go find a new room.
**Loris Marini:** I lo, I love what you think.
**Ashleigh Faith:** So like this, these are the ways that, that I tease things out from people is you're always gonna be dealing with folks that don't have a lot of patience, and that's fine. Just highlight the things that you need to highlight at the time. You need to highlight them. That's why I say early and often. If you start to figure things out early on, then you don't have to have these large conversations.
You're not gonna be wasting people's time. They all feel like they they're rowing in the same direction. And the worst is when you have a team either on the business side or the technology side that are so incredibly frustrated that they wanna quit. So I will say this is one of the biggest takeaways I've seen from this that I did not expect is this is helping with worker attrition too, because people get very frustrated and in today's market, especially with a lot of the technology, , you provide a lot of different jobs and you don't wanna lose these people.
And oftentimes the really cutting edge technology, like machine learning, like knowledge graph, like semantics of any sort, those are the things a lot of technologists, even though those folks sit in technology, your other technologists don't understand it all the time. And your business, people think it's magic.
So you really need to help those people specifically with what they're doing, because if they don't, Have the same understanding. If they don't have at least the same, what are we trying to achieve and how effective is it going to be? You're not gonna be effective with what you're doing and they're gonna get very frustrated.
And I have seen this. I can tell you from experience, there have been entire machine learning teams that were ready to quit until I stepped in. And not to say that I'm the special sauce, I'm not. I stepped in and started to do this facilitation with them and they started to understand on both sides of the coin and that's really what we're trying to achieve here is helping your people not feel so frustrated that they wanna leave
**Loris Marini:** In a world that is never been so divided. I think we just, we don't need their therapists. We need therapists in general.
**Ashleigh Faith:** in general.
**Loris Marini:** need
**Ashleigh Faith:** Agree on that?
**Loris Marini:** Yeah, that can Land both [00:40:00] hands and they're happy to do this, cross them bring folks together without feeling in any way that they are threatening them or accusing them of anything or judging them.
That situation you just described can escalate and I leave that first person. I've seen 50% of the engineering team and the startup I was working for completely vanish, within three months we lost 50% of the engineers because of a lack of that.
What we really needed was an engineering coach, someone that would come in and do exactly what you just described, but we didn't have that Yeah, so imagine the loss. I'm thinking about the business impact of this not only domain knowledge that is lost projects that don't have their owners anymore that can maintain them.
So all of those hours that you paid very at a very high rate. They're top dollar people. They're gone, right? . So you could, you should avoid that. So I think it should be a top priority for any team of any domain, especially in data, to have folks that bridge those gaps and and we need to train them.
So that's why I think. Perhaps me and what you already doing on your YouTube channel . But personally, I would love to be part of one of those dare therapy sessions. I don't know. It sounds like a bit of a dream
**Ashleigh Faith:** Oh, that actually would be fun to do it as a video where people can actually see how it runs. That would be fun. But the other thing though, just to add on, because I think it's incredibly important is a lot of people are still trying to figure out the whole machine learning AI stuff.
And I have seen this over and over and over again that they hire a machine learning engineer. Thinking that person will be able to manage themselves and the business and everything else because they're experts, A data therapist or somebody that will help them translate what they do to layperson terms for the business and understand how to translate what the business needs from the machine learning models is I say this very confidently.
This is one of the number one reasons your machine learning projects are failing a hundred. because you don't have that person that's translating between the two. And I see it over and over again. And a lot of companies, they wanna rush into machine learning. And I'm highlighting this spec specifically because I've seen it the most in machine learning.
I'm sure it happens in other disciplines as well. They need help desperately. And the machine learning engineer is oftentimes the not the one who's going to be able to do that. So you need to give them as.
**Loris Marini:** I want to hope that the future everyone in data. No matter where you are, is engineering, science, Analytics will develop part if not all of these skills that we
**Ashleigh Faith:** Mm-hmm.
**Loris Marini:** I still think that there is a strong case for a separate person to lead that, that exercise.
And the reason why I say that is, is a curriculum driven development which is an expression that I heard today from Joe Race in the podcaster data world Canada and Cocktails podcast with Team Gas Aspirin, Juan Secu. And he's right. I, a lot of the time in tech, especially in engineering, people want to be.
Heck, we are excited because we are on top of technology. We want to use technology to solve business problems, and we wanna stay on top, and technology changes so quickly. So what better way to stay on top than have real experiences? So whenever a new toy or new tool comes out, which is, the order of every week or every two weeks these days, you want to try and use it.
And that's understandable, but there has to be a counter force to that, and we can't expect these people to say, look, you know what? I'm not gonna make a decision to benefit my own career. I'm going to put the business above my own personal interests. It's a bit of topical, I think to expect that from individuals.
We all have agendas, right? We all have priorities. at good organizations is one where individual agendas are synchronized so that the individuals feel safe, they feel productive, they feel fulfilled, ideally, and everything is in sync as opposed to they're trying to go that way and the business is trying to go, the team is trying to go that way.
It's all part of the alignment. That's why hiring takes so, so much, right? But we can't expect an individual software or data engineer to say, I'm not gonna, I'm not gonna pursue new technology. I'm not gonna try and use it at work because that's not what the business needs right now. Ideally you would [00:45:00] want that, but I totally understand why a lot of data engineers feel like, you know what, if I have five, five tools on my resume, my, the chances of me getting a new gig after this one, Especially with, if you look at how long does it, one did engineer on average last on a roll as two years, so there's a high. You come and go, you get a job, and then technical depth raises to the stars and nobody can maintain it anymore. And you walk away. That's the story of data engine, of data teams. It's very rare to be part of a team that builds the foundation's. And they incrementally reap the benefits of what they do over and over.
And I don't know. I'm just talking about many different things. I don't know if I have one sentence to summarize this, but I feel like. Actually I do. My one sentence is there has to it's very beneficial to have someone else, someone that is not directly involved in doing curriculum or resume, dream and development and can see things more objectively
**Ashleigh Faith:** . Mm-hmm. And there's actually two tips I can give on how to do that well too, as that person and the first is always split the exploration for models. Again, talking machine learning specifically here compared to the implementation, always split them into two different features.
because, and this leads to the second tip and that first feature, you allow your engineers to pick three different ways to do it. And that gives them that, Oh, that's cool. I can play around with this new neural network.
**Loris Marini:** like a hackathon.
**Ashleigh Faith:** Yeah, it, so we wanna make sure that it's not completely open-ended, right?
You have very time boxed pieces and it allows the engineers to first get exposure and understand. Is this actually right for the for what we're trying to do? So I always say pick two to three, right? And I also, with my machine learning teams, we're constantly, and this is actually a good skillset for any machine learning person you're hiring, is constantly be looking at new papers, looking at new ways to do things.
I, we actually just keep a really simple spreadsheet where we talk about, here's the papers, here are the solutions, the problems they're trying to solve so that when we do find a problem, we. Go to that spreadsheet and say, Oh, we have a few things that we found that we can try for that. So that cuts down on that research time, allowing you to then experiment with something cool and something that is not as complicated.
For instance, heuristics, right? Like I know a lot of people that. In the machine learning space, they're like, No, I don't wanna do, simple heuristics when we can use this crazy, neural network to do it.
**Loris Marini:** it's reductive.
**Ashleigh Faith:** Yeah, but how much is it gonna cost to run that,
**Loris Marini:** And to maintain. Yeah,
**Ashleigh Faith:** And to maintain it to explain it and also AI explainability I think is incredibly important.
So if you can't explain how this thing is working to the business, and that doesn't mean you can't explain how neural networks work, you absolutely can't,
**Loris Marini:** Yeah. But then in a way that this books would get it.
**Ashleigh Faith:** And so that's really what you have to focus on is first having that approach to make time, to have your engineers play with something really cool, it gives 'em that exposure so that.
They can decide if it's a good solution to whatever next problem there is. And honestly, it might be the better solution. You just don't know. It's usually the simpler solution that works better though. And to your point, that's not as career building as something else. But this is a good way.
It's not a hackathon per se. It's really a literacy of what else is out there. Let's get some exposure to it and decide what is actually best for the business for this problem.
**Loris Marini:** So data translators, data therapists, data coaches many d. Aim for the same thing. Someone that takes the time to ask questions and slow things down and give people the space they need to be heard, to experiment and have fun. Because in the end, we're always there. We are also there to have fun and to together come up to a better understanding and feel safe about sharing ideas and with no judgment.
Sounds easy, right? Just read a book the book from Ashley Faith, which is assume is coming up soon, right? Ashley? No pressure
**Ashleigh Faith:** Oh Mo. Most of my stuff is on YouTube for all of this, but yeah, why not? Sure.
**Loris Marini:** No, definitely. I'm super keen to to follow up that idea. If you ever feel like. Hopping online, we could do a live session,
**Ashleigh Faith:** Yeah.
**Loris Marini:** on LinkedIn. And I was thinking with Ron Elman and [00:50:00] Ethan what's his family name? Ethan. Ethan. Ethan has been, Publishing a lot of interesting content on LinkedIn recently, and Ron we'd recorded a podcast together, is in the pipeline now, and he said, You know what, we should do a product, data, product design session live.
So we are going live soon. And we're trying to experiment. We have, we don't know what's gonna happen, to be honest. We just know we are three people very excited about data as a product, and we're trying, and we all have problems we need to solve. So we're gonna try and do, I'm going to try to be the facilitator in in that conversation.
Something that I haven't quite done before. It's always a
**Ashleigh Faith:** You'll do great at it. I'm sure.
**Loris Marini:** Yeah, it's always a one to one. So yes, send me resources. I can learn from you. I can only learn from you cuz you've done it in so many different contexts. And next time you do something like that, if you do it on your own, please let me know because I love to participate. Yeah.
**Ashleigh Faith:** Yeah, I will, We'll think about the logistics cause I think that would be beneficial for folks to, to see that happen. I always think like also I wanna do like a digital whiteboarding session because that's the other thing that's driving me crazy. This. Is everyone is saying everyone has to go back to work so we can have effective whiteboarding sessions.
I'm like, What do you think we've been doing the last two years, , we didn't have whiteboarding sessions. We have, and if you weren't, finding them effective, maybe you just didn't have the right tools or the right way to do it. I know that sounds. Terrible. There are very good reasons to have physical whiteboarding sessions, don't get me wrong.
But I think that some of them can be more effective than others. So giving some tips on whiteboarding since it's so prevalent in what we
**Loris Marini:** Look you definitely get one subscriber. I'm gonna be there in the audience, and we just need to figure out basically the top, the topic. I guess it's challenging because if you're part of the same team, you have the same, or. You wanna hope that you have the same business objective and kind of understanding or what the business is trying to do.
So it's a little bit easier to get that conversation going. Whereas if you have a bunch of people from different industries, different domains, different departments, getting them together to do a, their therapy session on they don't even know each other, like it can be
**Ashleigh Faith:** a that's why Yeah, you wanna make it intimate, small, right? And so getting reps from every. Area that you're gonna be working with. You probably don't need a giant therapy session like that. Start with the folks. It's almost like the Venn diagram, Who has the most overlap?
Have them, those data therapy sessions first, and you get farther and farther out, because then you'll have more and more people on the same page. Thick and help. And we all know, there's the water cooler conversations. There's, it starts to percolate through the rest of the team as you go.
So by the time you get to the end, everyone's Oh yeah, I heard from that we're gonna be, going this way. I was a little concerned. But then they explain their situ. It's almost like you're giving them the tools to also have their own little data therapy sessions. They don't realize they're doing it.
**Loris Marini:** Yeah.
**Ashleigh Faith:** So it's nice. It's nice when that happens.
**Loris Marini:** propagates. Yeah, definitely. So that's actually, this is a question I haven't asked. Is the size of the ideal size of a data therapy session? Based on what you said so far, or least what I understood so far, There is a big range. It could be a one on one session or it could be a one to many with many five or even more.
**Ashleigh Faith:** Yeah I would definitely say you wanna cap it at five to six. After that, you're going to get the leads. and then the people that are quiet. And it's, and I'll just say from personal experience, it's too hard to maintain good facilitation when there are more than five voices to be heard. In fact I've described it once you get past five, it's like herding cats on catnip.
**Loris Marini:** Yeah. Yeah.
**Ashleigh Faith:** go crazy in these things. That doesn't mean you can't have a larger data therapy session. It means that you need to have those smaller ones first so that you have more advocates on whichever way you're going. As a facilitator, you shouldn't be advocating for one direction or the other.
**Loris Marini:** But at least the intention, right? People advocate like they know why they're there.
**Ashleigh Faith:** Exactly.
**Loris Marini:** Yeah. Awesome.
**Ashleigh Faith:** they've been through it.
**Loris Marini:** Yeah. , Ashley Faith YouTube channel, actually, Ashley Faith knowledge graph and data. Absolutely subscribed to that. I couldn't recommend the more. There's one particular link episode that I wanna add to the show notes is with specifically on how IKEA does taxonomies and ontologies, which I think was fascinating.
And I have a long list of ideas that I think, Oh, I'm [00:55:00] gonna develop into some postal LinkedIn. I'll tag you and we'll we'll keep educating the audience. We need to because educating is. Mostly for me, very selfishly to learn, right? Cause when you're explaining something, you learn twice.
**Ashleigh Faith:** absolutely. That's why I do what I do. I wanna just help people do this more effectively. Get rid of all the mysticism and wizardry that people think that it is. Let's just talk about what it is in real life and move on. Get better.
**Loris Marini:** Ashley, you're LinkedIn as well. Is there any other place to follow your content and learn from you?
**Ashleigh Faith:** That's pretty much it. I try to dedicate myself to I'm very heavily on LinkedIn and YouTube. Outside of that, I do actually have a full time job, so I have to be careful how many other socials I'm on.
**Loris Marini:** Yeah. Yeah, definitely. Okay, fantastic ma. I'll make sure to add those in the show notes as well. And I just wanna say big thank you for educating me and everybody that has been listening to us, so super appreciate it.
**Ashleigh Faith:** Thank you