How Zalando used automation technology to create efficiency and improved candidate experience

Thank you, Johnny. So see if this works. All right, so interview scheduling overall is a topic which I will talk about today. So interview scheduling is a super important part of the recruiting process in many, many ways. But nobody likes the calendar, Tetris or like the email ping pong back and forth, which is required. Tech chief finds, find the right time slot for the right candidate to come in and in the end close the deal because. All right, so overall, interview scheduling is something which is like a fundamental part of recruitment. So when I worked on recruitment process outsourcing, I learned from many, many ways that interview scheduling and coordination in some cases takes up to 45% of the entire recruiting process. So that's what I will talk about today is like this, let's say, yeah, this area of talent acquisition recruitment, which is not the most attractive in many ways, but which is a super important part for us to automate and to optimize so very quickly about myself. So my name is and I'm heading talent acquisition at Zalando Overall. Just a quick overview about who Zalando is. So I already saw a few hands, but just a quickly overview, like who does not know Zalando? No hands. All right, so then don't have to tell too much about it. But overall, Zalando, like we are on a mission to become the starting point for fashion. And starting point for fashion means for us. So we want like we are already the number one fashion destination in Europe, but we want to become way more like the starting point for fashion and overall grow way more than we have done so far. So where are we today? So we are today active in 17 different countries and different markets. So we have around 2000 brands which are represented on our platform. We have like employees for more than 130 nations and overall are now to 16,000 employees. We're around 5 billion or a bit more than 5 billion, close to 6 billion in revenue and also the growing 20% year on year. So and with growth of 20% year on year, as 11 years old company which is already at around 6 billion in revenue, you can imagine the talent acquisition is like a massive topic for us. And for us it's not just about like how can we hire more, but how can we be more efficient, how can we do better talents to and then fulfill like our passion, our vision to become a starting point for fashion in Europe. So overall, let me take you through the journey which we have started in automation. So when I started at Zalando two years ago and what we did is we analyzed what are Our processes in talent acquisition, what are our processes in recruitment and how do we actually run our processes. So we could have said, oh, well, you're beautiful just as you are. But overall, the topic was they were not really customer focused and customer obsessed. Customer orientation and customer centricity is something which is deep in the DNA of Zalando. And so we started again looking at what are our processes and how can we actually optimize this. So we had a look at what are the moments that matters and where do we actually add value and what are all the other moments? So what are all the other steps in the process? And overall, for us, this was like a guiding decision to say everything which is a moment that matters for a candidate or for a hiring manager or for an interviewer and where we can add value with work, which we do. That's something where we want to double down on excellence and experience and actually dealing with this point in the process and everything else more or less, when it doesn't add value, when it is not a moment that matters for the people involved in the process. Let's automate it. Let's have a look at how can we actually optimize it, how we make it lean and how can we make, like, have systems in the end doing that? So that was like a very important point for us to understand what are actually the moments that matters for us and how can we deal with that? What we did then is we had a look, and that's like a short extract from, like, one area of our candidate experience survey. We originally have, like, this run in Qualtrics, and when we had a look at this, we saw that, like, scheduling of interviews is, for us, it takes around 30% of our, let's say, workload in talent acquisition. And then we had to look at, okay, well, how happy are people with interview scheduling? So we understood interview scheduling is something, when you have a look here, it's like it says, it's not important, but highly rated. So for us, that means, okay, well, how can we automate this? Because it's not a point which then really influences candid experience, but it's something which has to work properly and has to work properly at scale. Because just in Berlin, we have more than 12,000 on set interviews a year. So that overall, this was like a big scale which we wanted to automate. So then we thought, oh, no, nobody likes interview scheduling, right? But it's an important part of the process, so how can we actually deal with it? We had to look at what our coordinators actually do today and what do we actually want them to do in the future. So for us it was important. Interview coordination or like interview scheduling might not be like the most important topic in the entire process, but there are many, many things which actually the coordinators do on a day to day basis which add huge amount of value. When people come to on site interviews, make sure that they feel welcomed, make sure that like one interviewer is off site and they have to jump in right away to solve something or so on to in the end create memorable experiences for the candidates. But in the end don't sit there and just spend hours and hours and days in the end to do the interview tetris and find out when can we actually do that. So our goal is to have the interview coordinators becoming candidate experience owners and specialists so that they really own the candidate experience. And with that really focused not so much anymore on the manual and admin task around interview scheduling. So we ask ourselves the question, how can we automate this manual task overall while giving the coordinators the time to really create memorable experiences with candidates? And that was something which was for us a super important topic. So we went through the different process steps and analyzed what are actually the common pain points in the system or in the process. So the first one I already mentioned a few times, calendar tetras and calendar tetras like for us, which means when people come on site for on site interviews, they usually have somewhere between three and six on set interviews and one go. So that coordinator spend a lot of time opening up six different calendars and trying to find out when is the best time for them to like to come in and which interviewer has time when. Second one is manually updating several calendar invites as soon as there are any changes. So we're receiving candidate and interviewer availabilities which are not matching. They had to deal with different time zones. We have offices in different locations in different time zones. They had to find rooms, generating video conferences, links, find interview replacements and overall really manage the interviewer on the one hand, from a preference point of view, but also from a workload point of view to understand what are actually the things which we can give to an interviewer and which we can expect from an interviewer and what are the different things which you can also avoid. For example, when we interview lots of backend engineers, how can we avoid always same interviewers interviews with people and how can we actually make sure that we have a good balance and workload throughout this? What we did then is we did a lot of market research. So we had to try to understand what kind of tools out there, what kind of processes. So we thought about, okay, like there are chatbots out there, how can we actually do the email ping pong with chatbots? We had to look at automation solutions. And then we also had a look at, okay, how can we actually what other tools are out there. So we decided overall that to go for a pilot process or like a pilot in the end, which is for us a proof of concept. So what we decided on is that because we are using Greenhouse, is that we use the tool Goodtime. So we ran a pilot process, like pilot for several months with good time on like our tech hiring and overall, then in the end decided that like the, the process overall worked so well. So that we decided to go with this, with this tool and this partner. So in the beginning, which was interesting when, when we did like when we started our pilot, when we start our proof of concept, as we, in the beginning, like together with our provider, with good time, we did like an efficiency study. So in efficiency study, there were really like consultants coming in and sitting with like a clock beside our coordinators and measuring how much time does he actually take in different steps of the process and how long does it take when they get a task until they really finished it. And something which was really interesting is when we have a look at the results here we can see like the HR screening to schedule, like the manual task to schedule an HR screening took around 5 hours 90 minutes on average. So with the tool, after the tool implementation, they were able to do this in 55 seconds for a phone interview with a tech team or the hiring manager. We saw that this went down from 6 minutes 35 to 56 seconds. And even like an on site interview, we were able to schedule instead of 16 minutes within less than four minutes. And that was something which was super interesting for us to on the one hand understand, yes, we can be way faster. But then on the other hand, also when we have a look at the time distribution to schedule an onsite interview, we can see that there were like lots of different tools, There were lots of different task which they had to do. They spent 42% in the ATS. In Greenhouse, they spent 26.7% playing calendar Tetris. And in the end hopefully being successful with that and then also working with spreadsheets and so on. So overall for us it was clear that during the proof of concept, during the pilot, we saw that A, yeah, it's more efficient, we are faster, B we Already we can harmonize a process and also give our coordinators more time and so the end to do quality work. So overall we decided, yes, these results for us were very successful and very promising. So we rolled it out not only for tech recruitment, but we rolled out to the entire company. And overall at the moment we cover with that approximately three, three and a half thousand hires a year. And we are just about to roll it out also to the logistics sites for us. So how does it work overall? What we did is we connected Greenhouse with Good time and then overall our cool calendar integrations so that what we can see is that Greenhouse has in goodtime we have like an API so it shares job and candidate infos. We see the interview dates and times. And also good time is connected to our Google Calendar system. So it overall also exchanges everything so that the tool itself can schedule everything right away in our calendar system. How does it look like? Let's quickly have a look at like from a candidate point of view. That's in the end what they get. They get an email with like link, please get it yourself or we want to meet you. And then they have a very personalized, we're working always with on the design, a very personalized landing page. So they can do it via mobile, they can do it on their laptops. So they can do it right away within two minutes when they have a calendar available. So when they click on it, they can right away see which interviewers are available when. So they see right away which dates are possible for them. And they can also see, okay, well these are the dates where they can schedule interviews. In the end they get the confirmation and overall have scheduled themselves just about when they are fast, just about in 30 seconds, which really before that required a lot of emails, back and forth, confirmation and so on. And now the system sends confirmation and puts everything in the calendar of the interviewers right away. Talking about calendars, how does it look like? So for us, on the one hand we have a lot of, let's say calendar invites which are blocked for the different interviewers. And we have on the other hand the big calendar block which goes out to candidate. So that's how it looks like for the candidate side and then also for internal side. So what the tool does is it blocks all the different things which are before were manual. So somebody had to go in and block like from 10 to 10:30 and had to block from 11 to 12 and so on and so forth. But now the tool does everything and also sends out calendar invites to the candidates, right away. And something which I find really cool is candidates can reschedule without a human being having to be involved. Because the system, depending on our preferences, we have created a preference that candidates can reschedule themselves once and then after that the coordinator has to jump in. So they can actually say, oh well, oh, I just confirmed it, but I have some more of a dentist appointment, so I can't do that. I can go into the the system and just reschedule themselves. And the tool has all the calendar changes and everything else around that. So that's how it looks like. So something which we found was very, very helpful and interesting for us was when like we are hiring a lot in tech, we are hiring a lot of like engineers, product managers, researchers and so on. And what we always have had in the past is that we, when we want to test, for example, Java skills, we have like a long list of qualified and trained Java interviewers. And what is the tool does is actually you can tell the tool look like, let's just take like one interview, which is any interviewer who has like JavaScript tech. And with that the tool actually looks at like when is the next person available. So it doesn't have to wait for person A who's maybe on vacation next week, but it actually looks at who's available the earliest. And then like creates like an interview pipeline for this. But then also like what we can see here, you can also have fixed interviewers like the hiring manager or somebody from the team who's essential. But then overall we can also have in the end like long lists of qualified interviewers and the system then puts the different things together. What we can do is that we have also recruiting metrics so we can see for example which coordinator scheduled what. And that was something which was super helpful for us to really understand which coordinator is actually scheduling how much. So that on the one hand we're able, we are able to control workload way better. And on the other end we can also have a look at like, okay, well if somebody is very, very fast with the tool, how can we learn from this person to look at how can others adopt that? Then we can also have a look at like how far, how many interviewers does an interviewer do, right? So how many? Like for example, before that we had like a massive Google sheet list with like all the different names and the coordinators had to put in like, okay, well this person already did three interviews this month. This person only won. But now we have everything in one system. We can also like on a regular Basis, for example, we celebrate hiring heroes. So all the people who helped us and did the most interviews. And with that it just helps us with pulling one report out of it. Then also it gives us real time recommendations to say, look, where do we have enough interviewers? And why is the pipeline or the process slow down? Which is very helpful for us. But then overall, the tool is nice, right? The tool is a tool, but what are the outputs? And something which we did in the beginning is we actually spent a lot of time with good time together to understand what are actually, what is the baseline. So where are we coming from and what are we then measuring later on? And overall, what we said is that we have one in our candidate MPS survey, we have one question which says scheduling of interviews is seamless. And again, like for us, when it's not a moment that matters or we add value, it has to be seamless and ideally automated. So this was like one of the key questions for us. And we had already before that like 72% positive answers from candidates, but now we already moved up to 84%, so we implemented at the beginning of the year. So and then probably the remaining 16% are, I think, standard when, like scheduling is a bit more difficult, they have to fly in from different countries and so on. So it's not just like the tool of scheduling, but overall this was something which was like a key success indicator for us because with that also our candida NPS score has increased a lot. Also from an efficiency point of view we recently measured. So at the moment since we implemented the tool, we were able to schedule more than 30% more interviews with the same amount of coordinators. So that overall, for a return on investment point of view, we're able to prove that, yes, we can do more with the same amount of people. So our aspiration was not to reduce coordinators, but our aspiration was to give them more meaningful tasks and also to give them like a voice in the system and to create memorable experiences. While in the end, of course, the tool has to be efficient because the tool costs also money for us. So our experience so far is Based on around 30,000 interviews scheduled with the tool since beginning of the year. And we already see that like more than 90% of our coordinators are using it for all of their interviews. So very few exceptions where we don't use the tool. Time to schedule is also something which like in the past was never something which we really had a lot of data on. And what we see at the moment is that average time to schedule for us, from the moment that the coordinator starts the process until the interview is confirmed, at the moment it's 19.4 hours. And something which always surprises me is that 10% of the interviews get scheduled within 9.1 minutes. And I think that's something which is really because it's a mobile first platform. So when they get it they can just like with one or two clicks just to get it done, get it scheduled. 50% get scaled within three hours. And I think just thinking about the 3, 4, 5, 6 emails which are usually necessary to ping back and forth before you found the right slot, that for us was a big game changer. And then 80% gets scheduled within 23 hours. Of course there are always people who don't respond right away or so so that then like the last 20%, like that's something which goes up but that's. That were results for us. We were super helpful and interesting looking at the challenges. Now we are working with a new process with a new tool since nearly 10 months. And the big challenges which we had is on the one hand you have to get access to the interviewer's calendar because they have to share that if the tool cannot access the calendar, the tool cannot really schedule properly. So overall what we did is we created a standard centralized email which then our colleagues can share their calendars with so that then we have a so called TA coordinator@the lando.de and with the email address the interviewers can share the calendar. And so far with us we haven't seen any big pushbacks or so because they share it with a standardized email and we can then access the calendars and also work with them to make more efficient. So overall the big challenges are as usual, like without a tool is very, very common, complicated on set interviews with like multiple people, many like things going over different days. For example, like that's usually a challenge for all the systems because the system as we have set it up is mainly looking at like having a panel one after each other and with that really having like a very efficient process. And when we have for example someone coming in like four days in a row every day at 5pm, that's something where the system struggles a bit. But that's something which we for us don't have so often. Then the last piece was like the executive interviews, right? Because you don't get access to the CEO's calendar, doesn't matter what you do. And that's something which we have seen is as soon as the interviewers have an executive assistant or Team assistant who's doing the scheduling and where the calendar is just blocked completely, then the tool doesn't work. So we have to work with the executive assistants to make sure that when they put certain blockers in that they give the certain blockers a certain name, for example, interview block. So that the system knows that yes, I can over schedule this. But that's something which we have also solved so that even now, like today, all of our executive interviews also get scheduled with the tool. So the key learnings for us automation of interviews is partly possible. You can't automate everything and just trust your tool completely so you can do it faster, you can do it more efficient and overall it works very well. But in the end you still need the coordinators to doing the manual work and making sure that everything works. But their role has developed a lot from actually doing all the small tasks to overseeing the entire process and just jumping in when things are not working. Return on investment is positive and overall efficiency is measurable. So that was something which was super important for us. The more standardized process, the better for us. In tech recruitment, for example, we have a very standardized process. So we have for example, like everyone has to go through who's applying for a senior Java engineering role has to go through a certain tech screening and there is like a long list of interviewers who are qualified and trained for this. Then it works very well. Sourcing, screening and onsite interviews. Scheduling is possible for us. Like the sources, for example, they also use a tool to just like have an individual link so they send it out so that people can schedule themselves for a first conversation with the sourcer. And that's all something which have proven to be very successful. Key learning number two, in the sense measurable measure your baseline before implementation. It was something which was in the end like a big game changer for us. Because when we did in the beginning the efficiency study, we were equipped with data and we were equipped with like was really having proof that it works. And that was something which was also important for us for like the future investment decisions to extend motivated key users are important for success. I think that's something which has proven to be an absolute key success factor for us. So we have picked three coordinators who are trained on the tool intensively and who were the ambassadors for the tool. And they were so excited after a while that just their energy and their positivity around that has really impacted the entire team and in the end really created a lot of positive drive. And then something which really was a game changer for us was scheduling overnight as something which is really, really awesome. When colleagues leave in the evening and just like send out 20 interviews to schedule and arrive in the morning and 18 are scheduled out of the 20 and they don't have to do anything, then you have like two adoption from the coordinators because they really, really like it. So overall for us, happy candidate experience specialists, so they now really can concentrate on adding really, really good experiences and in the end really work efficiently as well with the tool. So thanks a lot. I hope it was insightful. I'm happy to answer any questions.