Honey Dev, Head of People and Culture at Edvoy, Chennai
Recruiter by heart. Builder by instinct. People leader by experience.
Eighteen years in HR. I started in recruitment, learned the business through hiring, then kept moving closer to the problems behind the hire. Eventually I started building the machinery I wished somebody had already built.
What I am looking for. A recruitment leadership or senior People role in a business growing faster than its infrastructure. Give me a blank page and something worth building. I care as much about the culture that grows around it.
I am a recruiter at heart. Everything else grew out of that.
Hiring is the fastest way to learn a business, because you cannot brief a good candidate on a job you do not understand.
I like the pace of it. The judgement. The fact that you are dealing with a person on one of the more consequential days of their year, whichever way it goes. I have never thought of recruitment as filling an empty seat. It is working out what the business actually needs, finding the person who can do it, then building an experience that makes them want to say yes.
That last part is where most hiring quietly falls apart. It is the part I have spent the most time fixing.
How I run a hire
Before anything else
Understand the job, not the vacancy
A brief written from a job description tends to describe the last person who did the role. I want to know what has to be true in twelve months for this hire to have worked.
Source
Know where your good people actually come from
Source of hire is not a reporting line item. It tells you which channels deserve your time next quarter, which is the difference between a pipeline you own and a pipeline you rent.
Screen
Look for signal, not keywords
Keyword matching filters out the career changers, the self taught people, the ones who did the work under a different title. Some of the best hires I have made would not survive a strict filter.
Interview
Make the experience the same for everyone
Structured, consistent, respectful of their time. When the interview quality depends on which manager a candidate draws, you are not measuring candidates. You are measuring your own inconsistency.
Decide
Make the call quickly, then say so
Most offers are lost in the silence between stages rather than at the offer itself. Speed is a candidate experience feature before it is an efficiency one.
Join
Keep the promise you made in the interview
The first ninety days decide whether the hire worked. Everything you said during the process is now a thing you either honour or quietly walk back.
What I watch and why
Time to fillWhere the process stalls, stage by stage
Source of hireWhich channels are actually producing quality
Cost per hireWhat each hire really costs the business
Offer to joinWhere the candidate journey breaks down
Quality of hireRated at three, six then twelve months
Recruiter productivityWhether the team can scale without losing judgement
A successful hire is not someone who joins. It is someone who becomes better because they worked with you.
I want the person I hire to leave, whenever that day comes, with more capability, more confidence, more perspective than they arrived with. That is the standard I hold my own team to as well.
TwoThe function I build
Then the vacancy stopped being the interesting problem.
Most of my career has been spent in businesses whose ambition ran ahead of their people infrastructure. That is not a complaint. It is the reason I know how to do this.
You cannot hire your way out of a broken manager, a performance framework nobody believes in or a payroll that gets it wrong twice a year. Once I could see that, the job widened. Teams, policies, performance, HR operations, analytics, culture and the technology holding it together.
At Edvoy I built the people function from forty people to more than four hundred and fifty across India, the United Kingdom, Bangladesh, the Middle East and Africa. One function, not twelve. I treat HR as something you engineer rather than something you administer, which is the only way one function covers twelve countries without falling apart.
I am a culture builder. Everything I build eventually comes back to one question. How do people feel about working here?
Where I work and what I do there
I leadRecruitment
Talent acquisition, leadership hiring, funnel design and the candidate experience wrapped around every decision.
I shapePeople and culture
Culture, engagement, employer brand, retention. It starts well before someone's first day.
I buildPeople functions
Business partnering models, performance frameworks, reward, policy and the operating rhythm underneath them.
I measurePeople analytics
Workforce planning, hiring funnels, attrition, quality of hire and the dashboards that make people decisions visible.
I connectHR systems
Operations, payroll, compliance and process design across multiple countries, with less manual work and clearer ownership.
I experimentAI and products
Workflows and tools built from problems I ran into myself, including HireMatch and Ellie.
ThreeThe evidence
Three times the number on the dashboard was not the problem.
Each of these follows the same shape, because it is the shape I use every time. Find out what the business believes, check it against the data, change the system rather than the target, then say honestly what I would do differently now.
Case one, recruitment
Twenty points of offer acceptance, without pushing recruiters to close harder.
Roughly one in three people we chose, who had also chosen us, were not turning up.
The situation
Offer to join sat at 68 per cent. Every lost offer meant restarting a search that had already taken weeks, so the real cost was hidden inside time to fill rather than showing up on its own line.
What leadership believed
That recruiters were not closing well enough. The suggested answer was more follow up and more pressure at the offer stage.
What the data said
Drop off was not concentrated at the offer. It was spread across the gaps in the process. Long silences between stages, inconsistent interview experiences, then a stretch between verbal offer and written paperwork long enough for a competing process to finish. Candidates were not rejecting the offer. They were drifting away before it arrived.
What I changed
Rebuilt the funnel around the candidate's waiting time rather than ours. Committed response windows at every stage. A consistent interview structure so the experience did not depend on which manager you drew. Compressed the gap between decision and documentation. Then I rang rejected and shortlisted candidates myself to find out what it had actually felt like.
The result
Offer to join rose from 68 to 88 per cent in two quarters. Time to fill came down 23 per cent as a side effect, because fewer searches had to be run twice.
What I would do differently
I measured quality of hire at three, six then twelve months, but only after the funnel work was already done. I should have set that baseline first, so I could prove the faster process had not quietly lowered the bar.
Offer to join, quarter on quarter
68 to 88%Offer to join, over two quarters
23% fasterTime to fill
8.5 / 10Quality of hire, rated at three, six then twelve months
Case two, retention
The attrition everybody was sure was about money.
One business unit was losing more than a third of its people a year while the rest of the company sat well below that.
The situation
Annualised attrition in the unit had reached 37 per cent. Exit interviews were being collected but nobody was reading them together, so every departure looked like an isolated personal decision.
What leadership believed
That we were being outbid. The proposed fix was an off cycle correction for the whole unit, which would have been expensive as well as permanent.
What the data said
Compensation was a factor but it was not the largest one. Leavers clustered by manager and by tenure band rather than by pay band, with the sharpest losses in the first year. That points at onboarding and management quality, not at the salary benchmark.
What I changed
Structured check ins at fixed points in the first year so problems surfaced while they were still fixable. A manager capability programme aimed at the specific teams the data pointed to. Targeted pay correction only where benchmarking genuinely justified it, rather than across the board.
The result
Attrition fell from 37 per cent to under 25 per cent within two quarters, at a fraction of the cost of the original proposal.
What I would do differently
I waited for the attrition number to get loud before I looked at the pattern underneath it. The manager and tenure signals were sitting in data I already had, months earlier. Now I read exits as a cohort every quarter rather than as individual events.
Annualised attrition, one business unit
Case three, scale
One people function, twelve countries, no second HR team.
Edvoy went from forty people to more than four hundred and fifty. The function had to grow without becoming twelve separate functions.
The situation
Hiring across India, the United Kingdom, Bangladesh, the Middle East and Africa. Different employment law, different payroll cycles, different notice periods, different expectations of what HR is even for.
What leadership believed
That each geography would need its own local HR setup, which is the usual answer as well as the expensive one.
What the data said
The genuinely local part was smaller than it looked. Statutory compliance, payroll mechanics and contracts had to be local. Hiring, performance, progression and how we treat people did not, so letting them fragment would have cost us a single culture across borders.
What I changed
Split the function deliberately. One shared spine for policy, performance, hiring standards and employee experience. A thin local layer for compliance and payroll, with named ownership per country instead of shared inboxes. Then automated the repetitive HR operations work so headcount in the function did not have to scale with headcount in the business.
The result
Payroll accuracy of 99.8 per cent across all twelve countries, with roughly 25 hours a month returned to the team through automation and AI tooling.
What I would do differently
I built the shared spine first then retrofitted the local compliance layer under it. Doing it the other way round would have caused less rework, because compliance is the part that cannot bend.
FourWhat I built
At some point, using the system stopped being enough.
Both of these started as problems I hit from the inside. I designed around the problem first then reached for the technology second.
HireMatch
Hiring software, live at hirematch.in
A hiring workspace built around what recruiters actually spend their day on. Writing the brief, finding people, reading CVs, preparing for interviews, keeping decisions moving.
Screenshot of HireMatch goes hereSave the image next to this file as hirematch.png
The job brief and candidate scoring view.
The problem
Recruiters spend most of their day on the work around hiring. Searching, formatting, chasing, coordinating. The judgement, which is the only part a machine cannot do, gets whatever time is left over.
How I thought about it
I have run that day for eighteen years, so I did not start from what AI can do. I started from where the hours go, then asked which of those hours a recruiter would miss if they disappeared. Almost none of them.
The design decision
Automate the work, keep the judgement. The AI clears friction from around the recruiter. It does not score people for you, because the moment a tool decides who is good, the recruiter stops reading carefully.
What it does now
One connected workflow from brief to CV to conversation, rather than another tool bolted onto the side of the last one.
Ellie
Employee listening, live demo
Employee listening built around the journey rather than the calendar. Ellie checks in at the moments that actually matter, lets people raise things safely, then turns what it hears into something a manager can act on.
Screenshot of Ellie goes hereSave the image next to this file as ellie.png
A day thirty check in, with the signals it produces.
The problem
The annual engagement survey tells you how people felt in the fortnight they filled it in. By the time the results are read, the person who was struggling in month two has usually either recovered or resigned.
How I thought about it
I started from Gallup's Q12, which is a tested way of thinking about engagement, then noticed the problem was never the questions. It was the timing. So I moved the thinking into the employee journey. Day 15, 30, 45, 60, 90, then year one and year three, with the questions shaped by the moment.
The design decision
A conversation, not a form. A form gets you a score. A conversation gets you the reason, which is the only thing anybody can act on. Ellie speaks like a colleague rather than an HR manual.
What it does now
Conversation becomes signal, signal becomes action. Someone can ask about a policy, flag a concern or simply say that something is not right, through a channel only HR sees.
FiveHow I decide
The numbers say what happened. These say why I made the call.
Three things I hold to, then three occasions where holding to them cost me something.
FirstDiagnose before you prescribe
Start with the business problem. The HR symptom is almost never the whole story. It is often not even the right story.
SecondBuild what is missing
If the right system does not exist, I would rather design it than write a paper about why we need it.
ThirdTell the truth early
Including about myself. I will tell you where I am weak before you find out the hard way.
Three decisions
The hire I argued against, then looked after.
Being right about the hire was never the point.
I was once asked to hire someone I did not believe was right for us. I raised my concerns and explained exactly why I thought the fit was wrong. I was overruled. The business wanted to move ahead.
So I did what I think a good HR leader does. I supported the decision, then gave him a genuinely fair run at it. Not a grudging one.
Three months later we let him go. The concerns I had raised had largely played out.
I could have said I told you so. I did not. I helped him find his next role instead.
Being right about the hire was never the point. Doing right by the person after it fell apart was.
Why I ring the candidates we rejected.
Culture does not start on day one. It starts the first time somebody hears from us.
I do not outsource candidate experience to a survey. I pick candidates at random, people we rejected as well as people we shortlisted, then ask them what it was actually like dealing with our recruiters and our hiring managers.
Three questions. Did we communicate clearly. Did the interview feel respectful. Would you apply here again.
The third one is the honest one, because it costs them nothing to answer truthfully and it tells me everything.
Sometimes what comes back is uncomfortable. That is precisely why I ask. It is also why I ask personally rather than through a form that lets everyone stay polite.
Culture is not built on day one. It starts the day you first speak to a candidate, including the ones you turn down.
Why I do not trust a metric that has not been explained.
The number is the signal. Finding the reason is the actual job.
I look past the count of roles filled. Time to fill, cost per hire, source of hire, offer acceptance, recruiter productivity and offer to join tell me where the system works and where it leaks.
But a metric on its own invites the wrong fix. Offer acceptance falls, so the instinct is to push recruiters to close harder. That treats the number as the problem.
Moving offer to join from 68 to 88 per cent had almost nothing to do with closing technique. It came from understanding why people were drifting away, then repairing the experience around the decision.
The metric is the signal. The real job is finding the reason behind it, then changing the system rather than the target.
SixThe record
Where I have done this.
Eighteen years across education technology, classifieds, software, media and services.
Oct 2020 – presentChennai
Edvoy
Head of People and Culture
Built the people function from scratch as the business went from 40 people to more than 450 across 12 countries. HRBP model, talent acquisition, performance, reward, compliance, analytics. Designed and shipped HireMatch and Ellie alongside the role.
Jan 2020 – May 2020Chennai
Facilio
Talent Acquisition Partner, leadership hiring
Worked directly with the CEO on leadership hiring across tech and product. A short tenure, cut off by COVID.
Dec 2015 – Jan 2020Chennai
Sulekha.com
Human Resources Senior Manager
A dual HRBP and talent acquisition role supporting 600 people in the corporate division. Scaled the design team from 6 to 20 in sixty days, which became the niche hiring playbook other units copied. Hiring efficiency up 40 per cent through ATS adoption. Attrition down 15 per cent.
Dec 2013 – Oct 2015Chennai
EROS International
Human Resources Deputy Manager
Leadership hiring including CTO, Head of HR and product roles, partnering with the CEO and function heads. Introduced structured onboarding and new hire feedback, which cut early attrition risk.
2008 – 2013Chennai and Mumbai
Helios and Matheson, iNine Technologies, Sasken Communications
Early career, recruitment and HR
Large scale recruitment for global clients across India, the US and Singapore. This is where the range came from. It is also where I learned to run a funnel properly.
MBA, Human Resources, Loyola College ChennaiAon certified, Performance ManagementPOSH certifiedGDPR certifiedVisionary Leader, 2024Allies of Women, 2025