Building a Data Academy: Presentation to Pittsburgh Chapter, Association for Talent Developemt

Editor's Notes

• #2: Objectives are on the next slide.
• #3: I’m going to use “data academy” as the concept here, of course it sounds cool, but what does it mean? We’ll get into that and then into four elements of establishing one -- First, how to sell it, understanding the market forces that are at work and what’s at stake. Then, designing it, focusing on what the learner needs and expects. We’ll look at implementing it, what formal and informal techniques to use. Finally, evaluating it, how can we measure the kind of ROI we are going to get from this program? If there are any questions, I will go ahead and get started… If there are any questions, I will go ahead and get started…
• #4: I am not a professional instructional designer or technical writer. I have stumbled on all of this haphazardly after starting a blog many years ago. Back then I had my first full time job and I was terrible at Excel and I was shocked at how much it was limiting what I could do at work. So I just kept blogging and learning and eventually started getting paid to do this stuff and really started expanding my horizons. Over nearly the last year and a half I have been focused full time on designing data analytics training programs, working with a lot of these bigger names in the space. So this is backed on that experience. So really where I am coming from is as a “subject matter expert” although I have had a lot of exposure to the instructional design side of things. And I think there is a need for this field of work, there is not a lot of literature on this, while there is a ton on instructional design there are things that make this area of it unique, so I will try to focus on those.
• #5: Just want to get a pulse check on where everyone is coming from. I put Technical Writers in the king’s spot here but I think we can all group around this idea of building a data academy for as I will show you this has become quite a trendy idea here. I want to look at how we can all work together in doing this. And I am hoping that I can learn from you with some of this too. If you have some tips to provide in this area of training programs and for technical topics in particular that is awesome. I didn’t get too much into the theories and concepts behind instructional design or training as I figured you probably know this already. I made is more specific to data analytics backed on my experience and market trends. My goal is that when you hear the term data academy, and I really think you will, that you will have the basic roadmap and vocabulary down.
• #6: So let’s jump in here, show what the market forces are, how things have really accelerated in the last year, what some of the marketing out there is, and what to be aware of as you are sizing this up. Then once you’ve made the market-scoping case for doing it, let’s look at some objections for establishing one. I’ll also provide a definition so it’s more clear what a data academy is and is not.
• #7: Here are a couple of screenshots I have found recently about this catchy idea of a “data academy.” You will see that these are both coming from banking in the UK and yes knowledge intensive industries like banking are definitely the right candidates for this kind of stuff. These screenshots are a little old, from 2019, but I haven’t updated them in the deck for a reason that you’ll see in a bit. But you are likely going to see this idea of a “data academy” in the news quite a lot across industries. [Animation] we can even think about Amazon’s upskilling initiative as a data academy, look here how nearly all these roles are data-intensive, some even have data right in the name! So this idea of data upskilling has been a big deal for a couple of years and that’s only accelerated.
• #8: Well all those initiatives were pre-pandemic so imagine what the demand is now. The digital transformation has accelerated and that often goes with data. Once of the interesting things with the pandemic is that it blurred the line between professional and personal interest and use of data, and we will get into that how that personal element is often undervalued when we think about data initiatives.
• #9: OK so I am hoping that lays the foundation and market context for doing one of these. So let’s say you are thinking about it and you start scoping the market. And a good place to start is, oh I don’t know, McKinsey. They wrote this article on the topic, again back in 2019. It’s a great article and I had it on the suggested reading list for this session, so check it out if you haven’t yet. It really highlights something to be aware of as you start scoping partners in this initative. Now I have here this data science hierarchy of needs, and it shows all the infrastructure that an organization needs to have in place to achieve AI. So there are a lot of antecedents here, and this hierarchy just shows all the raw data, let alone the culture and the talent needed to do it. AI sounds cool, it’s like this magical thing, everyone wants to do the magic without necessarily knowing all the work behind it, right? But I have a poll and some data on the next page that most organizations need to slow down before they can really implement AI to this extent in an organization.
• #10: Accenture and Qlik’s survey of 9,000 employees around the world meant to gauge how comfortable employees are using data in their roles.
• #11: 21% In other words, vast majority of employees really don’t feel comfortable using data. Expecting these employees to feel comfortable with AI may be a bit of a stretch given the overall hierarchy of needs. The Accenture report has some other interesting findings on the typical level of employee comfort with data. I’m going to switch over to other statistics on the next slide indicating how executives rate their organization’s use of data.
• #12: This is from information management consultancy NewVantage partner’s annual survey. Executives from 85 mostly blue-chip firms participated (Capital One, City of New York) Now, these executives admit that their organizations have a long way to go to use data better. Again, the majority of organizations aren’t quite at the place to embed AI across the organization. So we’re at a rare place where employees and executives agree about getting better at data? So, what do we do about that to develop the talent? Well I of course argue a data academy is the right approach, and I will offer you a more comprehensive definition in a second, but first I want to hear from you.
• #13: Now, knowing what you know about talent development in general and what you’ve learned about the data talent market in particular, and you wanted to implement some sort of talent development program for data in your organization, what would you anticipate the objections to be? As you think about it, I am going to offer a more substantive definition of the data academy, and then we’ll go into the objections that I’ve most often seen.
• #14: Now are you are thinking about I will give you some more details as to what I really mean by a data academy. Now for this “definition” I borrow from a couple of places, Josh Bersin on the capability academy and Gartner on data literacy. But a data academy should feel like a place people go. Now of course these days that is likely to mean virtual but that is fine. It should feel like something a little different than the ordinary, with its own norms and cultures, it offers a distinct experience, and that gets us into learning communities which we’ll get to in a bit. The data academy is a place for data literacy, this second line is Gartner’s definition of data literacy. Finally I like to emphasize here that there is something unique about each organization’s data academy. It should be able to deliver unique things to unique people and it should be seen as something relevant to people’s work and their lives. Now, that’s a lot to take in and a tall order. So understandably there are going to be misconceptions, objections, etc. Let’s go over the ones I see and maybe there are others you can think of.
• #15: You’ve probably heard this one before/not unique to data analytics, but one thing is the idea that users should just access the information as-needed through online platforms. Maybe even a paid learning source. These platforms aren’t inherently wrong but they are not doing enough to situate the academy within the unique context of the organization. If all you’re providing is undifferentiated training, well it’s a nice benefit and employees can get that anywhere. So this really isn’t doing enough to set the academy apart.
• #16: Now another common, maybe not an objection, but alternative approach is to establish a crack team of data scientists who just do all the heavy lifting for you, and you don’t have to worry about establishing talent development programs. The issue with this in some ways goes back to the hierarchy of needs, which is that an organization which isn’t data literate is not likely to have the infrastructure. This can be very expensive, because data scientists are expensive, and the turnover is going to be quite high when they find that they aren’t in an organization that can really support this. So this isn’t a problem that can be outsourced or centralized away, everyone needs to be engaged
• #17: So those are somewhat more philosophical arguments, let’s go directly to the costs. First column has the major cost sources of a data academy. First off, unless you have a pretty big organization you may need or want some help in setting up the program, eventually getting into the content and designing what it will look like. Well we’ll talk a little more about establishing the early team but it may be a significant enough amount of time for some early leaders that you’ll want to explicitly codify part of their FTE for it. Finally there may be some infrastructure costs, things like an LMS. You’ll want to think about early on how to meld with what L&D offers here. I think it’s important when you evaluate how much something costs to put in the context of alternative. A data academy might be expensive, but compared to what? What is the cost of no data academy? Well it’s no secret there’s a growing market for professional education these days. Now it’s easy to see that and think that employees can just self study or use your learning subscription to the org’s benefit, but the thing there is that your best and brightest will get a data academy on their own, they will enroll in a bootcamp and leave. And that concludes the section on the scoping and selling, any questions thus far?
• #18: Any questions on the scoping and selling? We’ll move into design & implementation next.
• #19: OK let’s get a little more into how to establish the academy to fullfil its purpose, how to take in the first members, and so forth.
• #20: Now the data academy is more than just training. So we want to design this in a way so people can actually be excited about it rather than just another box to check or something they feel forced to do. And to get that right it’s really worth understanding your learner’s needs. I mean it is data, so you may feel like establishing this would be purely quantitative, but it’s quite important to understand the qualitative needs and motivations of learners. At the same time, as an organization what gets measured gets managed, so you want to set this up so you have a clear established baseline, which allows you to project a target to define program success.
• #21: Now when you think about the data academy, as hinted at before you should think of it almost as a data community, and there is plenty of great literature out there on building communities, but here are some of the important questions to think about. A data academy should be unique to the circumstances of an organization and its staff so you want to consider what mission and values it holds. It’s really making explicit these goals of helping everyone make sense of their data, to feel more engaged at work and in life because of it, etc. Now, there are going to be individuals who use data in all sorts of way who you should consider. So you need to have some learning path for all of them. So think about the types of personas who you will serve, again there is plenty of great literature on how to do that. Finally, you’ll want to think about how to get buy in both from the top down and the bottom up. So that means having a sponsor in leadership, not just for the budget but the cultural implications of the data academy. Also getting members have some autonomy and can show that the academy is really meant to be a community. It’s not just more mandatory compliance training.
• #22: Now it’s a good idea to assess the skills of academy members, both to establish a baseline and to set the trajectory of the academy, and there are plenty of ways to do that… Some more quantitative methods might include asking for self-rating. You may want to get a sense for how your learners are currently occupied at work and what tasks are most time consuming, you’ll see this is a big deal for data. You may also want to develop a more holistic learning assessment to take inventory of learner skills. An asset that might come from this is a skills matrix to map required and desired skills for a team or project. This forms a great baseline to track the competencies. But the qualitative measures are easily overlooked, so make sure you take some effort to understand the motivations and goals That Accenture survey mentioned earlier found that one-third of respondents took at least one sick day due to data-related stress. It can be a very stressful part of someone’s job. Take the time to understand what is stressing out your employees and how this academy can help. Let’s look a bit deeper into what those motivations might be in this next slide. I have another poll for you.
• #23: Here are some learning assessments you might use, it seems pretty basic to just ask why you want to enroll but I think we really want to focus on the student’s motivations especially for such a highly technical field we want to dispel notions of impostor syndrome. Then here are some more broader questions, just about where the workflow is now and how our students see what they do with data. Finally here is a standard self-report kind of question, hey that’s fine but you know what that is pretty broad, this is like asking how good you are at basketball on a scale to 1 to 5, there are quite a lot of standards of talent for that, so it might be better to ratchet down what we are asking and even to add some context to those levels of knowledge, so let’s take a look at what that would be on the next slide. Now you want to lay some baseline. Again this is going to be different for different people. But in my opinion we can be dismissive of what the average analyst has to do, that they weren’t trained for it.
• #24: Now we are getting really focused on what we are trying to assess, and what the various levels of mastery are. So from here a clear picture emerges about the technical status of the learner, but I think it’s just as important to get the motivational and even emotional state of learners as well so we want to balance all of these factors.
• #25: Now this comes from a 2015 poll of 2,000 private sector workers, this question in particular looks at what motivates them to learn online. Now I know this just online learning, and you may have your own opinions or data on this, but if you were to guess here which do you find is the most common motivation for online learning?
• #26: A lot of it has to do with productivity, people don’t like spending all their time cleaning data, they want to automate things. So really keep this as a primary motivator, that this will boost productivity and make work more enjoyable. I’ve got a screenshot on the next slide that is almost an industry meme at this point.
• #27: Data is a messy business. This article from 2014 found that 50-80% of data scientist time is spent cleaning data. This number hasn’t changed much and it’s not just relevant to the most technical of your data users. Now, a little of this just comes with the territory of using data, it’s messy. At the same time, most people really have no training on how to do this. It can be frustrating and can be a big cause of churn for data professionals. So recognizing this and putting this motivation to be more efficient at the center of the data academy sets it up for great success.
• #28: Here’s another one that gets overlooked -- we may not think this matters but it does. With the right framework learners will really start to see data analysis as a mode of inquiry, a means of creativity. I really believe that data literacy is like “literary” literacy in that, when done right, people start to see it as a fun way to engage with their world. OK, so now that we’ve looked at the intake and motivation-setting, let’s get a little more concrete in terms of delivery methods.
• #29: For these next few slides I want to focus on more of the nuts and bolts of building this academy. We’ll look at what media to use, and how to design more engaging data literacy content. Those are more of the formal delivery elements of the academy. You also want to think about informally, how do you make something like a data academy part of the culture?
• #30: Now let’s look at how this information will be delivered and yes it’s easy to just use “blended learning” as a solve-all but let’s dig a bit into that could actually mean for teaching data. We should have some in-person sessions, or maybe Zoom, although I couldn’t help put a weary face there. Again, a data academy should feel like a distinct place to go, and this is a great way to establish a communal spirit of learning together. Then there is plenty of online study and there is a lot of great educational technology to teach data skills. Where things can really get interesting is by blending the academy’s curriculum with real-world applications. This can be great, employees can begin to look on some problems they have seen in their workflows with fresh eyes as a student and begin to propose some alternatives, there is a sense of immunity that people get when they are students that allows them to speak more candidly about their work. Finally given some of the pain points and scenarios that come out of such sessions a hackathon can be an opportunity to really crack open some of these things. This can be a very democratizing force. In any of these cases, the goal is to establish a place people to go build their data literacy, in the context of their role. So let’s talk a bit more about the informal mechanisms of the data academy as well.
• #31: All those formal channels are important but perhaps even more importantly is setting up a culture of sharing knowledge. This is another meme-worthy article from the WSJ. I remember reading it and it not sitting well with me. I’m a blogger so I know that hoarding knowledge is really the least effective technique in blogging, so why wouldn’t that be the same in an organization? Well I finally understood why, and that came from understanding the practice of knowledge management. 2018 article
• #32: I like the conceptualization of a data academy as a data schoolhouse because it makes it less formal, that people from all skill levels are meant to participate and help each other, and get something out of it. This idea of a community of practice from knowledge management is finding its place in the data literacy field and I am seeing some organizations blend these disciplines in delivering a data academy. If you’re not familiar, the idea is to establish a communal learning and knowledge sharing thing, so that people do have channels to share their knowledge, and it’s actually scalable and valuable, and not a nuisance to anyone. So you do have the channels and the cultural buyin to share that you are good at Ecxel.
• #33: Any questions on the design & implementation?
• #34: OK great, so we have buy-in, everybody’s happy with providing the academy, enrolling, etc. Now how can we actually establish some kind of ROI for this?
• #35: There are certainly some ways to quantify the ROI of a data academy from the revenue side in terms of offering more compelling and reliable services to customers, however I am going to focus on the expense side as I think this can be made a little more concrete. Earlier I suggested that at the establishment of the data academy to put together a skills matrix. This lets you uncover and track any gaps in your teams’ and organization’s data competencies. Ideally, you can continue to update this matrix over time and watch how matrix fills itself out and makes more capable and confident teams. Another victory to declare is with time savings. I have personally seen and been involved with creating solutions that used to take hours and have been reduced into fully automated processes that kick off at the click of a button. That is pretty indisputably a return on investment because ideally that is going to free up your analysts to do more productive stuff. These next kinds of benefits may come further down the road but are still pretty concrete. First, if you really start to enable these processes and this culture, you will begin to disband the data breadline and that will mean that your data teams will be happier with what they are doing, they will stick around longer. There will also be more morale, as people get to do what they were trained for. Now, how long until the data can support these improvements? As you’ve seen this is a big program touching technology, people, culture, you name it. Large programs like this may take a year or maybe more, so be aware of that as you put the evaluation plan in place.
• #36: Now I love technology and I wouldn’t exactly call myself a futurist but I think it’s very valuable to think many many steps ahead. So here are some things to think about as your academy grows and how it can be an asset in ways you might not have imagined.
• #37: Now these are all things that I’ve hinted at in the presentation but that are hard to pull off immediately. First as your academy grows you want to think about the sort of infrastructure you’ll need to scale it. At the beginning maybe you do just have a few workshops and skills matrices but over time you’ll have to find a way to disseminate this information more efficiently. You’ll also want to make sure that the data academy integrates with other L&D or knowledge management. We talked a bit about personas, and this is a great start. But as technology improves you may be able to drill down even deeper and offer personalized learning paths. So you don’t have to think about the data academy as this monolithic structure with pre-defined cohorts and a clear starting and stopping point but as a career-long learning tool to stop by at whenever you need a little boost. And related to this is the idea of microlearning, getting that boost of help right at the teaching moment. So when an individual is struggling for example with what chart to use, they can get help right then and there. Again, I wouldn’t expect a data academy to sweat this stuff at the beginning, it’s going to take some iteration and scaling before this is needed, but in the spirit of beginning with the end in mind, these are some trends and tactics to be aware of.
• #38: Now this idea really takes us to the furthest reaches of the data academy, but the possibility is exciting. It’s become pretty common in the tech world for proprietary software. I think this is really valuable because it gives talent some sense of the tools they’ll be using. There’s so much information asymmetry and it can be the difference between loving a job and sticking around, or just bailing. The same could be done with not just data tools but data training. An early example of this comes from Rackspace. They needed engineers so badly that they made this academy available to anyone for free and it was basically a recruitment generator for them. This program has been sold a couple of times and it’s a more traditional bootcamp, but it’s something to think about. You want to make your data culture transparent and you want people feeling comfortable with data maybe even before they get there. Anyway, our conclusion is next. https://www.tpr.org/technology-entrepreneurship/2021-04-16/codeup-buys-rackspace-cloud-academy
• #39: Before we part ways, I wanted to share some recommended resources for you, including a worksheet I made for this session, and if you have any final questions, that’s awesome, and I’ll share my contact information if anything comes up later.
• #40: Great conceptual overview of data literacy, this is one of those adjacent concepts. So you will dive more into different levels of data analysis, a deeper dive into what this term means, etc.
• #41: This is really a good book , the author established a Twitter learning community but a lot of it can branch out into an organization’s learning community.
• #42: This is a worksheet you get as part of attending today.
• #43: I also have more resource like this if you subscribe. We just talked about open sourcing tools and processes and I want this to be an open source data academy, give you all the learning objectives and other assets you would need to build a data academy.
• #44: Thanks for coming! Feel free to contact me anytime, find me on LinkedIn, I also write frequently on this stuff so check out my website too.
• #45: Thanks for coming! Feel free to contact me anytime, find me on LinkedIn, I also write frequently on this stuff so check out my website too. 2 mins