People Analytics: A Catalyst for Hyper Growth Recruiting – Manjuri Sinha, Zalando

thank you so much all right I'm not the last speaker before lunch so the pressure is not on me I hope let's see if we can get some energy into the room with the with our quick move on what's the land Oh is about to do [Applause] all right okay so anyone here has not heard of cylinder wow this is an awesome audience so I don't need to introduce the organization but I will talk a little about the journey just to set the context here that we are talking amongst organisations that have been into play for 40 years 50 years we heard Nestle talking here we heard Lloyds Bank talking here and I am representing an organization that is a small teeny-weeny ten-year-old baby at the moment yes we love fashion we are a fashion technology company we started at maybe one of the worst times in in the world in the economy when Lehman Brothers had filed for bankruptcy the whole world or the economics of the whole world was in recession and when Ali mentioned earlier that he ordered his shoes and he was waiting for it I said that there was a time that our founders used to actually take the parcels to the post office and would also pick up the parcels on returns from the post office and that's the early initial team that Sadr started solando to a grown-up in 2014 we filed our IPO and in 2012 already we were profitable in the dark region in 2014 we were profitable at the all the segments that we work in we started as a very simple online shop and if any of you remember the first test model was flip-flops and that's how solando is known even till today that we sell flip-flops and shoe it was not until 2010 that accessories and clothing were added to this segment and today we have become Europe's leading online fashion platform which has around three hundred and thousand three hundred thousand product choices I'm here actually preaching to the converted so I guess I don't need to talk more numbers we we work and sell in 17 countries across Europe close to 16,000 employees and cater to 2,000 brand partners across the world this is our global footprint where our fulfillment centers warehouses as well as tech hubs are based out of you can see that we also have a small presence in Bruna which is also close to Stockholm why am I here and what am I talking about and why am I here talking in Nordic people analytics summit well so as Allie earlier told that I I take care and lead the technology talent acquisition team for solando as you saw yes we are a fashion platform but fashion technology company and hence technology as is at the core of everything we do we've after filing for the IPO we're a small company but we are chasing big competitors like Amazon ASOS far fetches of the world and to fight with the competition and make our own niche we want to be the first stop for fashion we want to be the Netflix or the Spotify or fashion and this is to do this we need to star for technology teams and the that was why I wanted to set the context of why are we looking at hyper growth what does hyper growth actually mean we want to stuff a lot of people in a very short time to cater to our business goals to cater to our technology goals a panel that was here earlier talked about if you want to hire more people in your HR team what skills would you look for and there was a mention of I want a data scientist in the HR team and I was like oh my god the talent market is already so small for my technology recruiting what happens if the HR teams also started looking for data scientists like I mean maybe I'm out a job and we already have a lot of challenge in the market so what I'll be talking about is is a is a short story so keep in context and your old company still baby steps yes we I think and worded from Excel sheets three years ago when asked when I joined I was working in Accenture before I moved to solando so the first day I asked my boss can I see the knowledge portal can I look at via convey can I access reports that was my accent your habit she laughed at me left the room went away to another meeting and that was my start with solando from there to now for a lot of you sitting here you're experts you would feel that oh my god these are like baby steps such juvenile stuff but for an organization that was about 13,000 people then this was a challenge and how did we convert this challenge how did we crowdsource we heard this term earlier as well look at different teams we didn't have a people analytics team and we started this challenge how do we crowdsource capabilities how did we use and double down on what we have within solando already and what did we achieve through it so this is a slight story that I would like to tell about our challenge in 2018 so well what was our challenge as I said our in 2018 the business pose that we want to hire 1,000 engineers just 1000 engineers and I say Engineers yes we use a job family taxonomy here as well and job our engineering your family includes developers as well as data scientists researchers machine learning experts data analysts people who are working with the data Lake etc no data Lake is not a lake of data sorry bad joke I know for this forum what were our options when we got this challenge well we could have done what everyone has always done or what we've also done in the past hire more recruiters to ensure that we have enough resource capacity and load balance go for an RPO or what we call as recruiting process outsourcing hire a vendor or maybe deploy 10 to 12 head hunting agencies and spend a lot of solano money on getting these 1,000 engineers but well just as you saw in the video solando has always questions data school we've changed the market and we also believe in doing so within the organization we we have our fuckup Mondays so we believe in failing and learning from that and hence we wanted to look at a very different battle plan for this challenge that was presented to us for the first time ever we looked at something which was not not what the team or even me we were not used to doing this um and this was also because we had a new leadership team within pno there was a change of mindset there was a change of how we want to look at capabilities we wanted to target increasing our increasing our delivery so we looked at what was the rate of offer acceptance we wanted to increase that quarter on quarter from quarter one to quarter four we wanted to have a percentage increase in the number of offers that were accepted by candidates hence having more hires for that year we wanted to do this by looking and talking to numbers and what kind of numbers would we look at we would look at our pipeline quality that is how many candidates are coming in what is ideally the on-site to higher ratio what do you mean by how on-site to higher ratio so for every candidate that we bring in we have some on-site interviews and then if the person is successful we hire that person right so we want to look at if we if you want to hire a senior scholar engineer it takes us seven on sites to one higher and that would give us predictability and also the quality of the pipeline so we wanted to reduce the on sites to higher ratio the second bit is we wanted to look at time to hire that is what what is the present time to hire and what are the obstacles and how do we bring this time to hire down so that we can process more candidates and the third and very important for us because those are our customers that is candidate experience what was the status quo on candidate experience how can we increase this experience increase experience would mean more acceptance of offers as well as people who would be our ambassadors and refer other candidates in the future so what did we do in various steps and stages I would cover a couple of things I can't go into all the all the things that they tried and failed tried and failed but I will talk about certain activities which I we shook hands with technology yes we looked at what is within the organization we looked at what is outside the organization what I drew from C was also about talking about building a network looked at what was the network of vendors who were already approaching us what is relevant for us and what was the Londo famous for we looked at different teams and crowd sourced on what they could do I will talk about three activities and three things here number one is as an organization we have our business intelligence team we worked with customers when you go to your Zalando site or Zalando app you have something called as a recommendation engine working for you just like Netflix the recommendation engine or the recommendation of team that works with us in our in our sizing Department also looks at recommending what is what runs large what runs from also and so forth right so we do have a business intelligence team we have a recommendation team we have the technology within the organization what do they use and what are we not able to do we move from an excel sheet to an ATS about three years ago a proper 80s we use green house as an 80s green house would give us reports green house would give us a lot of information but it was all white noise this was not the white noise that our hiring managers a business wanted to hear this was not the white noise that was giving us a lot of information so we consulted the business intelligence team and rather looked at pushing this entire data through the redshift database and get full analysis data dump of how our candidate pipeline looks like this is our 1.0 and again referring back to Steve here I would say that do not look at the quality of the data at first go it is not a waterfall process it is always in improvement and we want to look at then improvising on this in 2.0 and 3.0 and so and so forth this was a first version to look at what is the candidate pipeline certain things that was very obvious from this candidate pipeline and the process was our filtered out so we have around 88% of applications are filtered out that means the recruiters who are sitting and screening applications are actually spending and wasting their time on looking at applications that do not even match our job descriptions and your profiles so that was the first area that we thought we would tackle yes numbers scare us we are from the HR field numbers scare us I admit numbers scare me but when numbers start talking and telling a story I like them I personally prefer them that's where we saw that ok this this is these these are the 88% filter out ratio how do I reduce this first of all time wastage and reduce the filter out ratio looking at that we tried to say all right what if we look at the job descriptions maybe what we are looking at what we are attracting and how we attracting is where the problem lies so how do we attract this we looked at our quality of job descriptions this is a freeware that is textio it's one example of looking at freeware that is available in the market you can also definitely go go ahead for a full subscription of textio but the freeware also does your basic job looks that neutrality looks at the length of the job description and in the end also looks at the conversion rate so you can see for this example we had a conversion rate increase from 16 percent to 25 percent in this altogether you can also get a conversion rate for cumulatively for all job positions that was the first effort to increase the quality of job wats then the second effort was to understand who and what are we targeting and this is how we need to understand how where do we get this data from not everything comes through our historical employee reports so we wanted to interview certain employees we redid some internal interviews we did some external interviews developed what we call as the Lando talent personas we also have this for our consumers and customers as customer personas and that's where we learnt the tactic from our brand marketing experts to develop talent personas where does our so if we have an engineering talent based in one of our logistics warehouses where does this engineering talent look for the next job what kind of websites is he or she looking into where they are usually toggle when they're using their free time based on this we now know where to target for a particular job family or a particular job when we're looking at certain channels and also how to message them what kind of messages are acceptable to this audience in this subtype of persona so we are not spending time and effort and money in sending messages into channels that do not give us any input or give us applications which we have to sit and filter out last but not the least we talked about in the cylinder video you saw that make employees have a sense of ownership and this is where we want to give each recruiter a sense of ownership to analyze how their jobs are performing where which country is giving them a better footprint which channel is giving them a better footprint and in the end then they have more control on whether to post on stackoverflow where the github is a better source where the linkedin is a better source it's working the job is working great from Germany why don't we just change the language of the job which we present in English to German and see we get more hits so and so forth so this is something that the recruiters get a get in control of so we saw data telling us the story here of a filter out ratio we saw that okay we need to look at a job description improve the quality of job description target properly and give control to the recruiters to also decide on which channels to use going further there was another information which was available in the MicroStrategy dashboard there's a lot of information but just in this screenshot there's something called as time to hire and time to start we were talking about this earlier layer that business and recruiters or business and HR sometimes talk different languages when you're sitting across a table business doesn't want usually doesn't want HR on the table because we don't stock the same language and it's about talking numbers talking manners talking how much amount of money would you need to spend or how many days would you need to have a particular bum on the seat to solve your problem that's what they want to hear and that's what we should be able to tell them so we wanted to look at our time to offer and time to hire we you also wanted to look at how do we reduce this and bring this down because a time to hire of 70 days was really harming our entire process was really harming the delivery that business wanted to cater to one of our challenges here when we looked at our time that was spent within the pipeline in the whole conversion funnel we saw that there was a lot of time which was spent in scheduling and waiting for interviews to be scheduled for a candidate sometimes a candidate would wait for three weeks to have an on-site interview scheduled bring the person over from another country have the on-site interview done and then finally think about another offer think about an offer by that time this candidate would have three or four offers from five startups that are sitting in Berlin or Amsterdam and he or she would have signed the contract already so how do we cater to this situation this is where we looked at this is where the combination comes in this is where we didn't look in-house we looked at something what was available in the market we looked at what other organizations similar to us are facing his challenges that that churn the number of candidates that we turned and we looked at certain examples of Airbnb Airbnb was using a scheduling tool called good time and we tried to do a proof-of-concept study with them last year to look at our efficiency study we have a team of coordinators interview schedulers who basically call the candidate ask for the time look at the interviewers calendars schedule the interview so on and so forth in the proof of concept efficiency study we figured that it was taking 22 minutes for a coordinator to schedule one set of on-site interviews for one candidate it was taking 45 minutes to one hour to reschedule one interview this was so if you imagine if you have five reschedules in a day that's all that this particular coordinator would be doing and wasting time on most of our 60% of the interviews were being rescheduled because of candidate rejections because they were there was a lot of back-and-forth they wanted to change the time later maybe they couldn't take a leave so on and so forth when we brought in good time good time is basically an AI based tool that matches calendars of interviewers and matches calendars of candidates it the interface also allows this is the interface that a candidate sees so basically the candidate himself or herself can know and pick their own time and slot for an interview doing this what we saw as a difference over the implementation of good time during the proof-of-concept and after the full rollout and implementation is the change in the execution so today we need 44 seconds to schedule one HR interview 44 seconds so that means the coordinator doesn't even need to do this recruiter can do this herself or himself second we saw 22 minutes for an on-site interview this was reduced to 4 to 7 minutes to schedule an on-site interview because our most of our schedules were coming in from candidates now candidates had the power to select their own time and slot the number of rejections a number of postponements from the candidates reduced by far and hence the rescheduling time also dropped in a sense then what what did the what did the coordinators do well the coordinators can now spend their time away from the laptop really guiding meeting candidates and enhancing our candidate experience and this overall helped us in reducing the time to bring a candidate on-site and hence reducing the time to offer overall it also created very good interviewer experience because in this we could really also drive it into the system that we cap our interviewing 22.5 interviews in a week and don't over utilize a particular interviewer just because he or she seems to be available so this is something which is automated coordinators today are only needed to have human intervention and escalations from the tool they also have cases where it's a very simple process and you don't have to schedule something with a vice-president or a senior vice president then this process can be automated and cater to the candidates even when people are sleeping or not in their working time zones then the last bit which is customer facing is candidate experience we wanted to enhance our candidate experience and that's where we looked in words again we have our customer NPS score so PR fashion tech company and our customer NPS score is one of the most important things that drive our drive our performance in business so that was customer NPS core this also called see NPS we just converted that into candidate NPS score use the same tools we use called Qualtrics for our Zalando customer NPS scores that gives us NPS for different countries fun fact Netherlands is actually our highest on NPS scores across Europe just a fun fact and we use the same tool this gives us an idea so you can see the the spread idly gives us an idea what should we work on when we ran all our candidates through the process whoever come whoever comes in and has an interaction with a human being that is the first step gets the set of questions whether they rejected or whether or are selected based on this we found certain areas that we needed to work on number one was speed which you saw he worked on earlier that was the time to hire number two was clarity again something we worked on based on the job description and improving the job description and number three was feedback Canada is told us number one that the feedback that was given to them was number one not actionable or not relevant in in number of situations we went back to our training team we looked in certain workshops we also went back looked and our hiring managers change the way we reject candidates especially when we have silver medalist candidates or who can be selected at a later point in time we have the hiring manager as well sit during the rejection call and offer a rejection call to the candidates and this made and interesting increase in our candidate experience altogether why is the Panda there whereas it's it's just a personal preference I like pandas and here it depicts a happy candidate as happy your panda is as good your acceptance rate would be what happened because of all this well this was these were the main activities that were driven we had a lot of other things going in the background well so what was the result we reduced our days to offer from 52 to 32 we were able to increase our candidate NPS score we started in January 2018 at a minus seven we ended up with a plus 33 at the end of November beginning of December I don't count anything end of December because everyone is on holiday so beginning of December and all in all we were also able to reduce our filter out ratio to a whopping 71% which was a major change from the 88% that we saw and overall we were able to increase by doing all these activities were able to increase our offer acceptance rate by 30% from first quarter to the fourth quarter in doing so at the end of the year we delivered around 700 hires our target was 1,000 and you would say you didn't achieve the highest yes of course so we changed our targets also during the year volatile business certain businesses decided to focus on other aspects and move the requirement to the next year we've worked along with the business and delivered on goal we were not we didn't see a headcount gap at the end of the year except maybe five or six some some places which we took carried over to the next year and we did all this by reducing our agency spent to less than 2% so to summarize and I think a lot of things would resonate with presentations that have that we've seen earlier is it's not always maybe important to get bigger so sometimes people always think let's hire more people let's get more resources the resource capacity in some times it's also important to get better and that's where our numbers help us second is double down on what works for us our branding works we know how to look at branding for our customers we know how to look at consumer data we know how to look and analyze consumer data we have tools that that do that tools that gave us consumer intelligence we wanted to use the same tools keeping it super simple things that the team can understand as we said unfortunately we don't have data scientists as recruiters at the moment I don't know how the headcount cost will actually align but that's for the comp invent team to think about but keeping it super simple something that the team can understand something that the team can translate to business something that the team can enhance on in the future as well and that's and in the end I think numbers can do as much you need people to deliver this and this is the team that delivered this for solando last year thank you [Applause]

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *