The Use Case Podcast - Storytelling about Panalyt with Daniel West
Panalyt CEO Daniel West was recently invited to speak on RecruitingDaily's The Use Case with William Tincup podcast, where Daniel and William discussed the main challenges faced by clients and prospects who approach Panalyt, along with a few high-value use cases and applications for People Analytics that the Panalyt team has recently been working on.
William Tincup: Today, we have Daniel West on, from Panalyt and we are going to be talking about both his company and his experience. Daniel has a fascinating background. He has been on both sides. He can talk about both sides of being a buyer and seller of HR Tech. With all of that backstory, Daniel please introduce both yourself and Panalyt.
Daniel J. West: Thanks very much for having me on William. I really appreciate it. My name is Daniel West. I am one of the co-founders and the CEO at Panalyt. The origin of the product really comes from my background. I was an in-house HR for over 20 years. I was the head of HR for Apple in Japan and Australia during the later iPod boom and the first iPhone boom and was then with Apple in the US, in Cupertino running HR for the US Sales operations and the Global Online Store. More recently, I was with Uber during the super-high growth phase. I was head of International HR where we went from a few hundred people to several thousand over a course of 18 months. I then did HR consulting, basically trying to replicate Uber Growth with a large number of high-growth startups. I was working for VCs and investors who would parachute me into their most recent investee company. So, I got a lot of exposure to super data-driven and very high-growth-driven organizations. I was in the position, particularly during my years of consulting, of buying the HR tech stack for those companies and trying to replicate what we had at Uber, which was a very very good investment in HR tech very early on, but I started to encounter a problem which has plagued me my whole career which is getting your hands on the data to support the business. And working for these very data-driven startups, managers who are used to having data on their fingertips to make commercial decisions, for them HR tends to be still a relative black box of data. We have this explosion in SaaS HR tools, but there has not really been, especially when we started Panalyt three and a half years ago, there weren’t tools on the market that would easily plug into all of these different SaaS HR tech tools that you’ve bought and extract the data and put it into useful and meaningful visualizations and so that’s really what we have built at Panalyt to fill that gap. What Panalyt does is that we connect to the APIs to all of your sources of people data, we consolidate it all in a custom-built data warehouse we build specifically to house and support people data and we built that data warehouse to deliver data very very fast to custom-built frontend where we have pre-designed dashboards to help managers understand their people data better and help them make practical people decisions on their day to day people-management challenges, That has been the core of the tool. What we have been doing in this last year during this whole COVID quarantine period with this explosion of people working from home, is that we have dug into another source of people data which is communication data- the data from GSuite and Slack, etc, pulling that communication data out to layer in how people connect to each other, the relationships that people have and the activities- how much time people are spending in meetings, on emails, etc- and melding that together with their people data to give managers an even richer view of how their team is responding to COVID. So, that's me and that's the Panalyt tool.
William Tincup: So when you say that I hear, for whatever reason, I hear BI for HR. Daniel J. West: Absolutely. It's BI for HR. It is the data pipelining, the warehousing, and the visualization. We don't think of ourselves as a competitor to anything like Tableau or Power BI where you have a data scientist who is building completely custom dashboards to the C Suite. We have pre-designed dashboards, and visualizations where you don't need a data scientist to build this stuff out, you don't need a lot of tech resources. Again, having been the buyer and having been on the inside, I never had any access to tech resources. No business leader who is short of developers for the front end of their actual product is going to give me any dev time, and again I'm never going to get any data science time from the front end of the business. So, you need a tool that is out of the box, already set up, to just absorb the data from any of your sources and deliver it to you in pre-designed visualizations visuals that will work, that will do all the calculations correctly and get all of your business partners and line managers on the same page as to what the headcount is, the attrition, gender balance, gender diversity, all of those fundamentals. We have extrapolated from that to start building more predictive analytics, attrition risk, connecting the correlations between performance, engagement, compensation, etc.
William Tincup: So because you are rolling everything up that will show you current and historical. Does Panalyt currently or do you see it in the road map to do forecasting? Daniel J. West: So, there's some simple forecasting around. We can forecast your headcount, your staff costs based on historical trends. You can choose linear trends, you can choose a smarter model, if you have it connected to your recruitment tool, it will look at what your approved recs are, what salaries are attached to those kinds of recs and start to predict it - on that basis or it will just take a linear of the last twelve months of your salary growth and extrapolate that. So, that’s the simple forecasting and then in the more complex predictive analytics, we have started with attrition. We have a predictive model that will predict your attrition over the next six months, the individual risk that someone is going to leave over the next six months based on a basket of dozen different attributes that almost all of our clients have. William Tincup: Is that data inside all the different HR Systems or are you using external data like recommendations from LinkedIn or something like that? Daniel J. West: That's a great question. So, we made a decision early on that we were going to focus on verifiable data that the client already has. So, we are not pulling in data that may or may not be correct, we are not pulling in crowdsource data or LinkedIn data that you cant validate. We want to have the foundation data that we are using to be absolutely concrete from the client’s systems. We really focused on leveraging the data that an average client will have as far as possible. So, if you take something like, if we know who the manager is, it’s not just a predictive power of does that manager has a high attrition rate, therefore, his people have a high attrition risk. Its - whats that managers historical attrition based on a certain gender, age, tenure, roles? How many managers have you had in the last year? What’s that manager’s own risk profile? So even just a single data point we can extrapolate seven or eight other data points and we think that’s much more powerful than spending a lot of time to pull in data that we are not even sure is accurate.
William Tincup: Again, position by position looking at some of these different indicators. So for recruiters- if a recruiter hasn't logged into their ATS in a month, that's one of many indicators that maybe there's a problem on the horizon. I want to go backward real quick to both your Apple and Uber experiences. I know people have questions but they also have a fear of scale. So going from that, 100 or 10,000 whatever it is, you have done it a couple of times and you have seen it a couple of times, so what's your best advice for both TA and HR as they think about Scale?
Daniel J. West: That is the question I've been asked most frequently in my career.
When I came into the international part of the business, we were maybe 300 people outside the US, and by the time we left, we were 4,500 and that was in 18 months. The first piece of advice is, Don’t do that! No organization can scale that rapidly and do it well. Or at least if you are going to do it, then just accept that not everything is going to go smoothly. But my main advice would be to invest in technology early, believe in yourself, and believe that the company is going to scale that rapidly. Have a clear plan about what your scaling is expected to be and invest in technology ahead of time. We invested in Workday very early and also we hired our workday implementation manager, which is when we actually got the value out of Workday, but I think very few companies do actually get the full value out of a Workday implementation. Investing resources into integrating the tools and connecting them together allows you to understand the processes and workflows involved across the employee lifecycle from recruitment to onboarding, to the different talent management processes being used for ctive employees, to exits, and automating them as early as possible. That means you’re making tough decisions and I think that being brave to make those tough decisions, for example, you’re not going to get to invest in training and development, you’re not going to get to invest as much in business partnering. You’ve got to invest in the processes to support hiring and onboarding early on, and again that will lead to problems, but you have to make those tough choices. And from a tech perspective, something that we didn’t do at Uber and where Panalyt comes from is that we didn’t invest in the output of all those systems getting it into the hands of the people who had to make the day-to-day decisions. So we did end up going down sort of blind alleyways in terms of massively over-hiring in one section but ignoring the other because we weren’t in touch with what the candidate pool was or where the business was going or where we were losing people.
William Tincup: I have heard that before. The phrase is “flying blind” Daniel J. West: Yeah! You're moving so fast and if you don’t have people or systems pushing the data to you or making it very easily accessible, you don’t have the time to stop or look around.
William Tincup: So, not semantics but is that Data insights right? Daniel J. West: Oh yes! Absolutely, it’s the meaning of the data.
William Tincup: Yeah. I hope when you’re saying Data, I believe that you’re really transferring it over into -” I want to get you the insight as fast as possible so that you can make informed, data-driven decisions”.
Daniel J. West: Yes. You are absolutely right.
There are these two concepts in People Analytics that I think are starting to emerge as the strategic avenues you can go down; this idea of the Push strategy where a small group of data scientists or specialists who are there thinking about “what does the business need to know so we’re going to do a project and then push the insights out to the business” and where it falls down, and where it definitely fell down at Uber where we had some of those team members but they didn’t have any mechanism to push those results out. There was no mechanism to get it out to the hundreds of managers that needed it.
Then, there’s that Pull strategy which is "Let's put the data out there in the interface where people who actually are running the business, have their hands on the business, they can derive their insights if we can make the data accessible to them, in an easy to understand accessible way", and obviously, that is exactly what we are trying to do at Panalyt is democratize the data to enable that pull strategy to work. So, in some cases, I am talking about delivering insights, but in some cases, I am actually talking about delivering data in an easy-to-understand, interrogatable way. The manager on the ground will have a hypothesis about why he is losing people. He or she should be able to interrogate the data and explore the hypothesis very quickly and derive his/her own insights. That is what we are trying to do with Panalyt.
William Tincup: Do you feel that in HR or TA, do you think that we are terrified about dirty data?
Daniel J. West: Yeah. I think we are way too over-conservative about data. I think we have been trapped in this idea of reporting. I think that’s very true for every TA leader I have worked with. They have come up through the trenches, and they’ve been beaten over the head that they need to be reporting, reporting, reporting, and if there is one data point wrong, they get hauled over the coals for it, and I think that leaders and HR on the ground need to embrace the idea that Insights & Analytics isn't the same as Reporting. Reporting needs to be accurate but Analytics doesn’t need absolutely accurate data. You need broad sweeping data set so that you can understand the trends and derive insights from the trends. It doesn’t matter if two out of a hundred people haven’t got their manager entered correctly or they don’t have the start date entered correctly. You want to get that stuff right. Don’t get me wrong, You want to get that stuff right, but it doesn’t mean that you cannot get the insights if those couple of data points are wrong and I think we are way too scared of the data being unclean and we don’t even start the data analytics journey. The other thought of this is that sunlight is a great disinfectant and your data will never become clean until you start using the data and understand where it’s wrong and fix it and see the result of it.
William Tincup: “I don’t want to start because I’m scared that it’s dirty…” Daniel J. West: Yeah! Yeah! “Because it’s too dirty to show anyone so I’m not even going to look at it, but if I don’t look at it, I can pretend that it's clean”
William Tincup: It’s like my closet. No! Different. Totally different.
Okay, so, a couple of things- As a practitioner, but also now as a vendor partner selling into your former peers - What are some common misconceptions about People Analytics that your team usually comes across?
Daniel J. West: I think the point about “I don't have enough data” or “My data isn't clean” is absolutely the biggest one. This idea that I don't have enough data- we really try and get our client past that if you are paying your people, you have enough to get started with. Your payroll data is much richer than you think it is. So, we always start with that. The second part is - “We’ve got too many other projects going on. We can’t think about Analytics because next year we’re going to have a new HRIS, let’s wait until then” or “We’re about to change our ATS, let’s wait until then” and the thing is that you are just delaying your ability to start educating yourselves and your stakeholders on understanding this data. People Analytics tools, whether it’s Panalyt or any of our emerging competitors, are built to plug into any tool. So, it doesn’t matter if you switch your HRIS in six months, it doesn’t matter to us, we will just switch the API. You’re delaying your journey of educating yourself on the business on leveraging analytics and democratizing the data. That is an error because every day you delay is the day you don’t get back and you’re always going to have another project and there’s always going to be another tool on the horizon.
William Tincup: It’s like firefighting. There’s another fire right now, burning and one would be burning tomorrow as well. Daniel J. West: Yeah! Absolutely. I think that people don’t have any problem admitting or understanding that if they had our tool in place, they would be able to fight their fires or even avoid their fires so much more effectively but they struggle to prioritize correctly about saying- “Let’s get these Firefighting tools in place first”. They want to have room to breathe. I think another thing about People Analytics is, I mean what’s so different about buying a PA tool vs buying another ATS or another HRIS, is that it’ll probably be the first time they’re buying it. We’re almost never replacing an existing People Analytics tool. Around 99% were being brought in to replace nothing or to replace a bad use of power BI or misuse of Tableau. It is not a tool designed for doing what they’re doing with it. Also, they don’t know how to make that choice and I think that also gets in the way.
William Tincup: This might sound like a trick question but it’s not meant to be that way. Who should own People Analytics? The reason I ask this is kind of a backdrop of just discussions I have had in the past with people from Workforce Analytics is that it’s kind of a combination of Operations, Finance, and HR, and everyone has some type of interest but who actually owns it. Daniel J. West: I don’t think that there is a right answer. It depends massively on the organization, the size of the organization, and the complexity. We don’t believe that there is a massive difference in industries particularly in how they use People Analytics but I do think that there is a massive difference in employee makeup. So, if you have a large group of retail workers, logistic workers, and field workers, then that is a very different employee makeup than an organization like Uber for example, or Twitter or organizations where there are 80% developers and everyone is in office.
Having a large group of retail and field workers changes the dynamic tremendously and things like that should help you to understand where the responsibility for the People Analytics should be sitting and how close to the business it should be. Fundamentally, I think it should be whoever has got the passion and drive to take on this democratizing journey and who has got the right leverage to say to the business that you really should be getting your own access to the data and having the drive to push it out. I think it’s got to be driven by who understands the opportunity best and so I think I'm not particularly dogmatic about which team should own it. I think it’s got to be driven by knowledge, experience, and confidence.
William Tincup: RIght. In some of the WorkForce Analytics, you see cross-functional teams come together. Finance learns from HR. HR learns from Ops. Ops learn from the Data Science team. No one has all the answers or you can look at the data and discuss what you see.
Daniel J. West: That’s exactly what’s been missing. So few companies have got a platform where you can all look at the same data at the same place. Where you can put People data together with Finance data and Commercial data in a single place and look at it all together.
William Tincup: And the larger the company, the more complex it’s going to be. I have had a lot of payroll providers as friends and clients in the past and it’s like a simple question from the CEO like “how many employees do we have?” and they'd say 700,000 employees after looking at 18 different payroll systems. That’s actually a hard question.
Daniel J. West: and the one answer CEOs hate is that “ that depends on what you mean by an employee”. They hate that.
William Tincup: “Hey can I get a week?” Daniel J. West: Well, they hate that as well.
William Tincup: But to be able to roll that up and go, like currently, as of this moment, here's where we are at. Daniel J. West: Yeah. It's exactly the same as between business units and, lot of my experiences in operating internationally, so between what a country like Brazil thinks its 500 employees, but on your data, it's 200 employees. So, where's that difference of 300, and very often it’s not a small difference.
William Tincup: Right! So, now you unlock this idea that you've got contractors. Are they actually employees?
Daniel J. West: Yeah, exactly. They might be contractors and they might be on a different system may be because locally you need to pay them differently.
William Tincup: Oh Yeah! That's right. Daniel J. West: And the pain the HR Business partners face is not necessarily understanding that you can understand it but you know that every time you want to run a report or build a piece of analysis for your clients, you have to extract the data from those different places manually, you need to put the data together manually, and everyone’s got their 5 or 10 items little checklist of all the things they know they need to change in the data every single time to make it slightly more correct. Then you put it together and put it into PowerPoint, etc. We’re trying to help clients go through that just once. Just do that once with Panalyt. Just once, and you won't have to do it ever again. It’ll just open up correctly in Panalyt.
William Tincup: Years ago I read this book, Lean Analytics and it came out of the Lean Startup movement. It was basically about how every business, unique to that business and probably unique to where they are in their maturity, has one metric that everything watersheds from. For Airbnb, at one point it was the number of uploaded photos. So, they could look at the health of the business any day just by looking at it. When you look at People analytics, is there a magic bullet, anyone metric that they should be tracking?
Daniel J. West: I think that’s a great question. It does come up. We have put some effort recently into what we call the People Balance Sheet. We're essentially taking the Finance balance sheet format and putting in all the critical metrics for HR, but you, your team, and your company will select those metrics and you will have them in an easy-to-understand balance sheet format. I think that is the start of the process of understanding what your critical metric is. When you get all the stakeholders viewing the same balance sheet, you will find some sort of gravity around one or two numbers, one or two lines that everyone starts focusing on. But, what I have noticed is that the number that everyone is focusing on will start to change between maybe Q1 and Q2 or Q3 and Q4, and people will start to focus on some other number. So, I would shy away from advising clients that there is one key metric, but I would advise clients that there are a limited number of metrics that you can get everyone to focus on. Understand what is driving your People challenges today and identify the couple metrics that help people be aware of and understand that issue but then develop the quality of all of your metrics so that once you have solved that challenge, you can move on to the next metric. What we advise in the sales process is not to try to boil the ocean, because people say “once I have got an Analytics tool, I can do this and I can do that and I can do that too” and we say that Yes you can but…”
William Tincup: “Yes you can, but the question isn't if you can, but if you should.”
It’s funny a hundred years ago I did an HR Conference in Miami, and one of the bits that I did, there were probably 70 people in the room and I said, I'm going to give you some time and you can discuss at your own tables, but I am going to go around the room and I’m going to ask that if you could only measure one thing, what would you measure and why? and you can give some context about the industry and company size but nothing else, and you know 70 people, 85 different answers. But, one of the things that we got which I thought was fascinating was “Regrettable Turnover”. So, not just turnover but, “Regrettable”. Daniel J. West: Ah! Yeah. That's a great one.
William Tincup: Okay. Two final questions.
One is, buying a People Analytics tool, buying Panalyt, let’s just make it specific, what are the questions that you love hearing from practitioners, that you think they should be asking of Panalyt? Daniel J. West: The questions that we like to hear reveal that they are thinking about Analytics the right way. So when we get the question of how do I solve this particular issue, when they have got a particular challenge and it’s like this thing is on us and how do we use data to solve this. That’s obviously a great opportunity. Also, it's nice when there's an openness to our advice. When we get questions like, what other data point could we be gathering to move the needle and actually it is interesting that you mentioned the Regret Attrition because that’s often the one I go to like, “Are you measuring Regret Attrition?” and the answer is invariable No. There's still only a minority of teams that are measuring that, and I would go like, “look if you could gather one more data point, and if you go over the last six months exits and just do regret and non-regret, implementing Panalyt, you are going to see so much more power from the Attrition Analysis.
William Tincup: That's right. The word ‘turnover’ to most HR people is like a bad word, and I am like, “No, Not really”. Trees in a forest die. It’s okay. Daniel J. West: Yeah. You have to start distinguishing between positive turnover and negative turnover.
William Tincup: That's right. Daniel J. West: You have to really focus on that first three months and the first-year turnover because that indicates a whole other set of issues. So, those two questions - how can I solve this issue and what other data could I be collecting- really show the sincereness.
William Tincup: I love them both, but I love the second question more because then they are thinking of you as a Partner. Daniel J. West: Yeah.
William Tincup: You know what I mean. It's like saying You've done this, and you see this all the time. What are we not measuring? What are we not thinking about? What's around the corner? Daniel J. West: Our whole sell is about that partnering approach and discovering what they are going through and them getting to see us as partners. One of my co-founders heads our Japan business and she is ex-Google People analytics and was at Google for years in Japan and the US. So, between her Google experience and my Apple & Uber experience, we cover a lot of great brand names that people want to hear from.
William Tincup: Which is good. All three of those companies figured it out on some level. So, great representation. Last question and I know this is a question you get asked often - you've built business cases and done cost-benefit analysis for the CHROs or the CFOs or anybody else you’ve dealt with, but how do you render or talk to them about the ROI of People Analytics?
Daniel J. West: The buying process for HR has always been difficult. It’s somewhat easier for TA but in general, it’s always difficult for TA and HR to show our ROI. I think the whole buying process is difficult and it has just gotten harder and harder over the last few years when you start factoring in the technical hurdles and the data privacy hurdles that you have to jump over just to get to the point of even starting to talk about the commercial advantages.
How we advise our clients to build the ROI around Panalyt, there's a couple of really obvious low-hanging fruits, so obviously, Attrition is a key one. If you can understand Attrition and leverage our Attrition Prediction tool you can start to lower your Attrition, and there are calculation methodologies to understand what attrition is costing you and so if we can reduce your Attrition, then obviously there's an immediate cost-saving. We also have a tool that starts to understand the connection between your recruitment behaviors and the first-year success of employees. So, who the hiring manager is, what the source is, who the recruiter is, how many interviews they've had, what the average interview score is? All those inputs, correlating with first-year performance and first-year attrition. So, again if you can show that I can improve your first-year performance and help you to improve your recruitment processes so that instead of losing 10% of your new hires, you're only losing 2% of new hires, there's an immediate cost-saving there. So, we have helped them with those kinds of obvious cases. Beyond that, it’s back to that question of what is their immediate pain point. If you can identify what the immediate concern is of the CFO or the CEO, then you can tailor your analysis to that. For CFOs, it’s very often, especially for high-growth startups, about spiraling unpredicted staff costs. The CFO wants to have a better predictive model on how staff costs are growing and more awareness on how salaries are increasing for new hires and how that's going to impact Q1 actual costs and Q2 actual costs, and we have a forecasting tool for that.
William Tincup: In general, CFOs hate variable costs. Daniel J. West: They hate unforeseeable increases. If you can show that as an HR team, you’re aware that everyone you hire is impacting their cost model but also that I can start to show you predictions on how we've already seen an X% increase over the last X number of months in our salaries, therefore what you have budgeted for Q2 is going to get blown out because of this trend. But not just telling the CFO that, but also telling him that he will have his own access to it and he can just open it up and see it for himself. Also, what frustrates CFOs is that HR is often a black box to them and if you can open up that black box in a secure and confidential way and give them the data they actually need, then you’re making a real friend there.
William Tincup: And also, they might see something that you can't see. Daniel J. West: Yeah. Absolutely.
William Tincup: I lied and said that was the last question but I need to ask you about the quality of hire. I get asked a ton about quality of hire as a metric, are you being asked about the quality of hire?
Daniel J. West: Yeah, that's specifically what we built our Recruitment Outcomes tool for. We got a fairly specific definition of quality of hire. We are looking for that first-year outcome. So, within the first year, are you a high performer, low performer, top quarter, or second quarter performer? Did you get a pay increase, did you get a promotion, did you quit. If we are connected to their communication data, we can show you that in that first year, are people becoming engaged in the company, are people building relationships inside the organizations? Do you see them as high performers? Are they still isolated even after 6 months or after 1 year at the company? Some of that is impacted by onboarding, but you can correlate it back to who the hiring manager was, how many interviews they had? Who was in the interview process? Who was essentially selecting the person and who was selling them on what working at the company is like and therefore setting the right expectation or wrong expectation. It's about the candidate quality but the quality of the hiring process that's going to produce that good outcome vs negative outcome.
William Tincup: You used that metaphor about light being the antiseptic. I really liked that, because eventually one of the things that you are trying to solve is - bad managers. Daniel J. West: Yeah. Bad Managers, bad processes. Also, you can have some really good managers who are just really bad at hiring. It's about understanding in what way you are a bad manager and can we help you to be a better one, instead of just going about saying that you are a bad manager.
What's interesting about attrition prediction is that you can identify managers who are, for example, really bad at managing people who are remote vs they're great at managing people who are in the same office as them. That's now changed in this Covid period, but it is important to understand the strengths and weaknesses of managers, and that's something that HR has been very very poor at because the data wasn't immediately available and it was just too hard to do.
William Tincup: Yeah. We paint with the same brush. Daniel J. West: Yeah, absolutely. You might know that the data is in there somewhere but you know you don’t have the skills to go and get it and you have like a million other things to do, but if you just look at one chart and it’s telling you, then you can do something about it.
William Tincup: I love it.
Daniel, we can talk for hours. Thank you so much for coming on The Use Case Podcast. I absolutely appreciate you. I love what you are doing at Panalyt.
Daniel J. West: Thank you, You too William, what you’ve been doing for years now and if I can just take a minute, several years ago, you were doing an annual review of the HR Tech Market and I found some of your work, about 4 years ago, when I was putting together my first Investment pitch for Panalyt, and I definitely used some of your writing in my very first pitch deck which got us the initial funding for Panalyt, back in the day. I have followed you since then, and I appreciate all the work that you've done to raise up this profession that we are in.
William Tincup: And I appreciate what you're doing and thanks for all the compliments and the kind words. Thanks to everyone listening to The Use Case Podcast. Until next time.
Panalyt bridges the People-Data Gap, enabling real-time, uniform access to relevant people data, reports and insights for CxOs, HR and business managers.
People data, including employee interactions and connections is combined with business data empowering businesses to leapfrog to data-driven decision making, eliminating bias and improving engagement, sales effectiveness, productivity and, as a result, business performance.
Interested in a further discussion on how People Analytics and Relational Analytics can help you drive an improved employee and business outcomes? Book a 30-minute discovery call with our Panalyt co-founders to learn more!