Utilizing Data in Recruitment Decisions

You hire great people. They perform well. Then they leave. So you dig into the data. You run the analysis. And you find something weird: performance doesn't predict retention. Your top performers leave at the same rate as everyone else. Maybe higher. So you conclude that performance doesn't matter for retention. Time to focus on culture fit, compensation, growth opportunities—the stuff that "really" drives whether people stay. 𝗡𝗼𝘁 𝘀𝗼 𝗳𝗮𝘀𝘁. You're looking at a restricted range. And this is one of the most common ways organizational data lies to you. Here's what happened: you filtered for high performers when you hired them. Now everyone in your dataset is a high performer (or close enough). 𝗧𝗵𝗲 𝘃𝗮𝗿𝗶𝗮𝗻𝗰𝗲 𝗶𝗻 𝗽𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗶𝘀 𝗴𝗼𝗻𝗲—𝘆𝗼𝘂 𝘀𝗲𝗹𝗲𝗰𝘁𝗲𝗱 𝗶𝘁 𝗮𝘄𝗮𝘆. When you remove variance in one variable, other variables fill the gap. Now the things that *do* vary—loss of a favorite manager, a competing offer, a shitty project assignment—become the dominant predictors. Not because they matter more in some absolute sense, but because they're the only things left that can explain differences. It's not that performance doesn't matter. It's that you've built a team where performance *can't* explain differences because there aren't enough differences to explain. This is restriction of range. And it's everywhere in organizational data. Why does interview performance barely predict job success? Partly because you only hired people who interviewed well. The people who would have bombed? They're not in your dataset (because you didn't hire them). You're correlating interview scores with job performance among people who all cleared the same bar. Why do employee engagement scores seem unrelated to team output? Partly because low-engagement people already quit. 𝗬𝗼𝘂'𝗿𝗲 𝗼𝗻𝗹𝘆 𝗺𝗲𝗮𝘀𝘂𝗿𝗶𝗻𝗴 𝗲𝗻𝗴𝗮𝗴𝗲𝗺𝗲𝗻𝘁 𝗮𝗺𝗼𝗻𝗴 𝘀𝘂𝗿𝘃𝗶𝘃𝗼𝗿𝘀. The pattern that "doesn't exist" in your data existed in the selection process that created your data. Here's the thing: before you conclude that "X doesn't predict Y," ask a different question first. Did I filter on X before I measured Y? If yes, you're not seeing the relationship. You're seeing what's left over after the relationship already did its work.

The CEO was about to cut the recruiting team in half.  Until the Head of HR changed one slide in the board deck.  Headcount was approved in 120 seconds. Leadership had already decided. Recruiting was "overhead." $18,000 per hire and 67-day cycles looked indefensible. One Director of Talent asked for five minutes before the vote. She pulled up the original deck. "Cost per hire is $18,000. Time to fill runs 67 days. Recruiting spend totals $1.2M annually." Board members saw exactly what they expected to see. To the board, it looked like pure cost. Then she pulled up a second version. "Engineering seat vacancy costs $4,200 per day in lost output. Before our team joined, average vacancy length was 94 days. Today it's 67 days. Those 27 fewer days across 48 engineering hires at $4,200 per day equals $5.4M in recovered productivity." Nobody had done that math before. "Mis-hire rate was 22% before. Currently it's 11%. Average cost of a mis-hire is $180,000. That 11% reduction across 48 hires equals 5 prevented mis-hires and $900,000 in avoided waste." Their CFO closed his laptop. "Our $1.2M recruiting spend returned $6.3M in measurable value. We achieved a 5.25x multiplier." Without hesitation, executives called for a vote. Headcount was approved in full. Eight minutes later, the meeting ended. Walking out, that Director of Talent called her HRBP. "They were going to cut us in half." "Tell me everything." "I translated recruiting metrics into business outcomes and they approved full headcount." HR has the data to prove financial impact. Most teams just report it in the wrong language. Leaders don't fund "time to fill improved." Boards fund "$5.4M in recovered productivity." Once the board saw it that way, $1.2M in recruiting spend became $6.3M in value.

Most HR leaders would hate me for saying this, but 90% of hiring metrics are useless. You don't need a dashboard with 47 KPIs. Here’s 7 numbers that actually predict whether your hiring is working: 1. Quality Applications Track how many candidates meet minimum qualifications versus total applicants. If you're getting 200 applications but only 10 are qualified, your job postings or employer brand need work. Quality beats quantity every time. 2. Time to Fill Days from requisition to accepted offer. Every day a role stays open costs productivity and team morale. Track by role type to identify bottlenecks…is sourcing slow? Interview scheduling? Decision-making? 3. Interview-to-Offer Ratio What percentage of interviewed candidates receive offers? If you're interviewing 20 people to make one offer, your screening process is broken. This reveals whether your pre-interview assessments actually work. 4. Offer Acceptance Rate What percentage of your offers get accepted? Low acceptance rates signal problems with compensation, candidate experience, or employer brand. Track by seniority level to see where you're losing top talent. 5. 90-Day Retention What percentage of new hires are still engaged and performing after 90 days? Early turnover is expensive and usually preventable. This metric reveals misalignment between expectations and reality. 6. Hiring Manager Satisfaction How do managers rate the candidates you deliver and the hiring process? Your internal customers' satisfaction predicts whether hiring best practices will stick. Low scores mean misaligned expectations. 7. Cost Per Hire All-in recruiting costs divided by hires made. Include recruiter time, tools, assessments, and external fees. Understanding true cost-per-hire enables better resource allocation and ROI discussions. TAKEAWAY: Most hiring teams measure activity instead of outcomes. These 7 metrics focus on quality, efficiency, and long-term success. Track what matters, improve what you measure.

Workforce planning has always been an incredibly complex and difficult task. Despite valiant efforts to improve these models, they have remained relatively static and simplistic, relying predominantly on small teams crunching data or on predictions from the hiring manager community. In an ideal world, we would shift from a static, once-a-year exercise to a dynamic, more proactive model. We would stop reacting to what's happening now and start anticipating what's likely to happen next. Last week, I had the pleasure of spending time with our enterprise data and analytics team, a group that services over 800 customers. The most exciting topic we discussed was three pilots we're running with customers right now that aim to make this a reality: using a digital twin for work planning. It works by connecting vast amounts of external market data with a company's many internal data sources, some they typically wouldn't consider, such as ERP, CRM (sales), LMS, and Time and Attendance systems. This allows us to run scenarios and model future talent needs. Here’s a concrete example: By analyzing Salesforce, HRIS, and ATS data, we can predict that when multiple prospect opportunities reach a specific stage in our customer’s sales cycle, there is a high likelihood of winning at least one of them. We can then analyze the consistent skill sets across all of those prospect opportunities, allowing us to confidently and proactively start a recruitment process for those skills. The goal being that we have candidates at the final stages of the process, before an official requisition has been raised, positively impacting time to hire. We’ve also been able to replicate a similar model based on website sales activity. The question to ask is: what data is generated in what system that allows you to get ahead of the hiring process today. 

Most recruiters brag about one number: reqs closed. That’s like measuring the health of a company by how many invoices it sent. Necessary, but nowhere near sufficient. You don’t get promoted because you filled seats. You get promoted because you created value. The recruiter of the future doesn’t say, “We closed 25 roles last quarter.” They say: → Time to Productivity → How fast did our new hires start delivering ROI? → Quality of Hire → Did they become top-quartile performers or just warm bodies? → Retention → Are we building teams that last, or running a revolving door? These aren’t HR stats. These are business metrics that are directly tied to revenue, margin, and shareholder value. Here’s the thing: opinion is cheap. Data is expensive. Walk into a hiring meeting with gut feel and you’ll get a polite nod. Walk in with proof that referral hires outperform LinkedIn sourcing by 30% and suddenly you’re not “just a recruiter. You’re an operator. Data isn’t about dashboards. It’s about power. It’s about credibility. It’s about telling the story that links recruiting to the bottom line. The shift is simple: stop counting roles filled. Start reporting on enterprise value created. Because recruiters who measure the right things don’t just fill jobs—they fuel growth. Question for you: What’s one metric beyond reqs closed you’re tracking that earns you a seat at the table?

Explaining HR data to non-HR business leaders - a simple model. I recently met with the executive leadership team of a large company exploring how to use technology to become a more skills-focused organization. This conversation required me to explain how skills data is related to other forms of HR data used to measure employee performance, qualifications, and potential. To help with this, I quickly created a model based on the fundamental psychology of human behavior that seemed to do the trick. There are four basic kinds of data used to manage talent. 1) Attribute data that measure people's qualifications and potential based on their past experiences combined with their personal interests, abilities and preferences. 2) Skills data that measure what people know how to do based what they have learned from their past experiences. 3) Behavioral data that measure how people have performed in their role based on either observing or tracking their actions at work. 4) Goal data that measure what people have achieved through their actions at work, either directly as individuals or collectively as members of a group. While any HR process might use all these types of data to some degree, in general: -Recruiting focuses on the first two types of data (1 & 2) to identify, screen and select candidates based on qualifications and skills. -Coaching focuses on the middle two types of data(2 & 3) to help people develop their performance and potential through building skills and managing behaviors. -Performance management focuses on the last two types of data (3 & 4) to evaluate and reward people based on achieving the right things the right way in their roles. Part of me thought this is a bit like explaining the primary colors to someone who wants to paint their house. But then I realized, just like not every homeowner thinks like an artist, not every leader thinks like a psychologist. Apparently, when it comes to explaining HR data to business leaders having extensive training in the measurement and prediction of human behavior does have some benefits.

Most companies still look at people data in silos. HRIS for headcount and attrition. Surveys for engagement. Exit interviews for why people leave. Policies for what’s supposed to happen. Each of those tells a clear story on its own. But the real story usually lives in the intersections. Here are three hotspots you can only see when you’re analyzing patterns across multiple data sources: (1) Structural barriers pushing your strongest people out. This often shows up as a highly engaged and high performing subset of employees with elevated attrition. If you’re only looking at engagement or performance, this group looks like a strength. If you’re only looking at attrition, they look like a retention issue. Put it together and the pattern is more specific. People who are motivated and capable are choosing to leave. That usually points to constraints in the system. Career paths that don’t translate into real mobility. Managers who develop talent but can’t retain them. Job scope or compensation that lags performance. (2) A breakdown between hiring quality and on-the-job success. Recruiting data says you’re bringing in strong candidates. But early tenure data shows lower performance or higher attrition. That gap is rarely about sourcing or selection. It tends to show up in onboarding, role clarity, or manager capability in the first 90 days. You only see it when you connect candidate quality signals with post-hire outcomes. (3) Policies that exist, but don’t translate to consistent experiences. On paper, your company offers flexibility, growth, and inclusive practices. At a high level, survey data might even support that. But when you layer in team-level outcomes and usage patterns, variance shows up. Some managers turn policies into real day-to-day experiences. Others don’t. The policy isn’t the issue. Manager behavior is. None of these hotspots shows up clearly if you stay inside a single system. And most importantly, these aren’t just interesting insights. They tell you where to act and what to fix. Looking across systems doesn’t just give you more data. It gives you a clearer map of where performance is being unlocked or constrained.

As recruiters, we’re not just matchmakers—we’re strategic partners who drive business impact. One of the most powerful tools at our disposal? Data. Here’s how leveraging data can help recruiters influence decisions and shape outcomes: Identifying Top Talent Trends: By analyzing sourcing metrics, application rates, and candidate engagement, we can spot patterns and proactively adjust our strategies to attract the best talent. Improving Diversity & Inclusion: Data helps us track representation across pipelines, identify gaps, and set measurable goals for building more inclusive teams. Optimizing the Hiring Process: Metrics like time-to-fill, candidate drop-off points, and interview feedback scores allow us to streamline processes and enhance the candidate experience. Influencing Stakeholders: Presenting clear, data-driven insights to hiring managers and leadership builds trust and helps drive alignment on hiring priorities and resource allocation. Measuring Impact: Tracking post-hire success and retention rates ensures we’re not just filling roles, but making quality hires that contribute to long-term business goals. Data isn’t just numbers—it’s a narrative. When we use it thoughtfully, we can influence decisions, advocate for candidates, and elevate the entire recruiting function. How are you using data to make an impact in your recruiting practice? Let’s share ideas! #Recruiting #TalentAcquisition #DataDriven #HR #Hiring #RecruiterLife

Recruiting teams today track almost everything. Open rates, replies, placements, time to hire, even how many messages were sent in a week. Having that visibility is great, but we’ve seen many teams still feel stuck even when the numbers look good. It’s not that the data is wrong. It’s that it doesn’t always show what to do next. At Pin, we try to focus more on what sits behind the numbers. When we start seeing shifts in candidate engagement or outreach performance, that’s where we learn the most. It shows what kinds of roles are catching attention, which outreach messages land better, and where things tend to slow down. Once you can see those patterns clearly, you can make adjustments that actually improve how people hire and how candidates experience the process. Good data should lead to better decisions. It should guide where you spend your time and help you understand what’s working instead of just reporting what happened. If it doesn’t move how you think or act, it’s just another chart on a dashboard.