No Company Can Tell You When Someone Is About to Leave. That Is Not a Technology Gap. It Is an Architecture Gap.

Experience Milwaukee founder Steve Glynn on why every organization he talks to is asking who is about to leave and nobody has a real answer, what hiring systems built on credential logic are systematically missing, and why rewarding failure is not a cultural aspiration but the specific organizational behavior that determines whether capable people stay…

·

21 min read

Experience Milwaukee founder Steve Glynn on why every organization he talks to is asking who is about to leave and nobody has a real answer, what hiring systems built on credential logic are systematically missing, and why rewarding failure is not a cultural aspiration but the specific organizational behavior that determines whether capable people stay or go.

The most consistent finding from hundreds of conversations with organizations across industries, sizes, and geographies is not that employers do not care about retaining their people. It is that by the time they learn someone is leaving, the decision has already been made and the conversation that follows is, at best, awkward and, at worst, irreversible. The signal that would have enabled a different outcome was present weeks or months earlier, in the behavioral patterns of a person who had already begun mentally and emotionally departing from the organization before they ever said a word to their manager. The systems of record that organizations rely on to manage their workforces were not designed to capture that signal. They were designed to record the transaction that follows it.

The talent and workforce technology market has produced an impressive set of tools for the moments that organizations can see coming: hiring platforms for acquisitions, onboarding systems for first days, performance management tools for annual cycles. What it has not produced, at any meaningful scale, is the infrastructure for the moments that matter most and arrive without announcement: the shift in an employee’s behavioral patterns that precedes a resignation, the credential-filtering hiring process that systematically excludes the candidates who would have become top performers, and the cultural environment that says it rewards risk while punishing every person who actually takes one. These are the gaps where the most expensive workforce outcomes are generated, and they are the gaps that no transactional platform was designed to close.

Steve Glynn has spent nearly two decades studying what makes communities and organizations attract, develop, and keep talented people, or fail to. As the founder of Experience Milwaukee and the builder of Sprinklr, a community platform he grew to the largest meetup in Wisconsin before selling it, he has observed these dynamics from the outside as an ecosystem builder and from the inside as a founder who learned firsthand what the behavioral patterns of an organization that is retaining and developing people actually look like when they are working and when they are not. His conversation here covers five arguments that HR tech platform operators and employers need to hear about the gap between what their organizations claim to want and what their actual systems and behaviors are producing.

About the Guest

Steve Glynn is the founder of Experience Milwaukee, a media platform focused on the people and organizations building Milwaukee’s tech and business future. He previously founded Spreenkler, a community organization he grew to become the largest meetup in Wisconsin before selling it. His City Tech Tour research examines how other cities build strong talent ecosystems, and his MKE in Six initiative is a six-pillar, six-year plan for Milwaukee’s innovation trajectory. His work spans ecosystem building, storytelling, and the kind of direct conversation with founders, operators, and HR practitioners that produces honest observations about what is actually happening in organizations rather than what they wish were happening.

Every Organization Is Asking Who Is Going to Leave. Nobody Has a System That Tells Them Before It Is Too Late.

The finding Steve reported from his conversations with organizations across Milwaukee and beyond is the clearest field validation of the post-hire blind spot that exists in any conversation in this series: not one organization he has talked to has described a reliable mechanism for identifying that an employee is moving toward departure before the employee has made the decision and communicated it. The pattern he described is universal and familiar to anyone who has managed a team: the employee tells their manager they are leaving, the manager is surprised, and what follows is a conversation that everyone knows has already been decided. The organization learns about the problem at the moment it has become irreversible, which is the exact definition of a system designed to record consequences rather than prevent them.

The organizational response to this pattern has been consistent and consistently inadequate: organizations invest in retention programs, competitive compensation packages, and culture initiatives designed to make staying more attractive, without ever building the infrastructure to identify which specific employees are at risk, when that risk is developing, and what specific conditions are driving it. These investments are not wasted, but they are deployed in the dark, directed at populations defined by demographics or role type rather than by the actual behavioral patterns that predict departure. An organization that spends significant resources on retention without the intelligence to know where attrition risk is concentrated is not managing retention. It is managing the average, and the average obscures the individual-level reality where every retention problem actually lives.

Steve’s observation that companies are simultaneously worried about who is going to leave and uncertain about where to invest given that uncertainty is the precise articulation of why behavioral workforce intelligence is not a feature but a foundation. An organization that does not know which of its employees is showing the stability pattern shifts that precede a resignation cannot make a rational investment decision about retention. It can only make average investments and hope the distribution of the risk happens to align with the distribution of its interventions. That is not workforce management. It is workforce lottery, and the odds are precisely as unfavorable as that description implies when the replacement cost of a single mid-level departure runs to one to two times annual salary.

“I don’t know anyone who’s able to spot anything early. What I see more is people say, oh yeah, I told my employer I’m leaving because I got a new offer or whatever the reason is. And by then it’s too late because it’s awkward.”

The infrastructure answer is what MustardHub provides through its behavioral workforce intelligence layer: continuous signal collection from the post-hire period that surfaces attrition risk and stability pattern shifts before they produce the resignation conversation that Steve described as the universal moment of too late. The behavioral patterns running through the post-hire blind spot are detectable weeks before they produce a formal signal in any system of record, and the difference between catching them at thirty days and catching them at ninety days is frequently the difference between an intervention that changes the outcome and a farewell conversation that does not. The organizations that build this infrastructure are not just reducing their attrition rates. They are transforming their understanding of their workforce from a lagging view of what has already happened into a leading view of what is about to happen if nothing changes.

For platform operators, the universality of Steve’s finding is both a market diagnosis and a product opportunity. Every organization in his network is experiencing the same problem. None of them have a product that solves it. The tools they are currently using were not designed for this purpose, and the tools being marketed to them as solutions are largely repackaging the same episodic measurement approaches that have been failing for two decades. The platform that builds continuous behavioral signal collection into its architecture, routes attrition risk to the managers and leaders who can act on it, and connects the leading indicators to the financial model that makes the investment case obvious, is building into a gap that the market has confirmed exists everywhere and that no current product is reliably closing.

Hiring Decisions Made Entirely on Credential Logic Are a Defense Mechanism, Not a Talent Strategy.

The research Steve described on hiring decision frameworks, the distinction between logic as a filter and person as a filter, is one of the more practically consequential analyses of why organizations consistently make expensive hiring mistakes despite having rigorous processes. The specific dynamic he identified is that logic-based hiring criteria function as a liability shield rather than a talent optimization tool: when a hire does not work out, the hiring manager who followed a defined logical process can point to their adherence to the criteria as evidence that the failure was not their fault. The criteria, rather than the judgment of the person applying them, absorbs the accountability for the outcome. This is individually rational and organizationally disastrous, because it produces a hiring system optimized for defensibility rather than for identifying the candidates who will actually perform.

The job description research Steve has been conducting across hundreds of postings, from conservative legacy organizations to technology companies at the frontier of AI development, reveals the same structural problem: every job description, regardless of its disclaimer language about flexibility and fit, is written to describe someone who has already been doing the job in a corporate setting with escalating responsibility for ten to fifteen years. This description does not fit the founder who built a company from zero to meaningful revenue. It does not fit the veteran who led teams in conditions that make most corporate management challenges look trivial. It does not fit the elite athlete who spent fifteen years developing the discipline, competitive orientation, and team-first behavioral patterns that produce high performance in complex environments. These are not edge cases. They are large populations of capable people that the credential-matching hiring architecture is systematically excluding, and every organization that has hired from them, as Steve described of his own hiring history at Sprinklr, has found them to be top performers.

The questions Steve proposed as replacements for or supplements to standard credential screening deserve to be taken seriously by every HR tech platform operator building in the hiring and talent acquisition space: have you taken an idea from zero to a functioning something; have you taken revenue from zero to a significant number; have you rallied a team around a vision and produced momentum toward an outcome. These questions are not soft alternatives to rigorous evaluation. They are behavioral assessments of exactly the capabilities that determine high performance in complex organizational environments, and they are capabilities that the standard credential filter was never designed to surface because they do not appear in the formatted history of someone who has spent their career inside the corporate credential ladder.

“How come you don’t ask, have you taken an idea from zero to a functioning something? Have you taken revenue from zero to some significant number? Have you rallied a team of misfits around the vision and gotten them all to get momentum on an idea or a company?”

The post-hire behavioral intelligence connection is in what happens when the hiring decision is based on credential matching rather than capability matching. A person who was selected because their resume fit the template but whose actual behavioral orientation does not align with what the role requires will show distinct stability pattern signatures in the post-hire period: early disengagement indicators, role-fit mismatch signals, and the behavioral precursors to the attrition that a better hiring decision would have prevented. MustardHub’s stability pattern data surfaces these patterns through continuous behavioral signal collection in the post-hire blind spot, giving organizations the ability to identify a hiring mismatch while there is still time to address it through role adjustment, additional support, or the honest conversation that acknowledges the misfit before it produces a resignation that everyone retrospectively agrees was predictable.

For platform operators building recruiting and talent assessment tools, the credential logic defense mechanism argument creates a specific product responsibility: if your platform makes it easier to hire defensibly rather than effectively, you are building a tool that produces compliance with the hiring process while systematically missing the candidates who would produce the best outcomes. Building the behavioral assessment questions Steve described into candidate evaluation workflows, alongside the capability to connect post-hire behavioral signal back to pre-hire assessment data, is the product architecture that closes the loop between the hiring decision and its actual consequences rather than treating them as separate transactions in separate systems.

Curious which of your own people are at risk?

Live insights on your workforce in two minutes.

See who's likely to quit

The Organizations That Keep Talented People Are the Ones Where Leadership Behavior Matches Leadership Rhetoric.

The leadership behavior observation Steve described, that what a leader does communicates what behaviors are expected far more clearly than what they say, is the behavioral foundation of every culture problem that organizations subsequently pay to diagnose and fix. A leader who sends emails at three in the morning is not just working late. They are communicating, with more authority than any company values statement, that working at three in the morning is what this organization expects and rewards. A leader who publicly endorses risk-taking while privately moving the people who took the risks that did not work out is not creating a culture of innovation. They are creating a culture where everyone understands that risk-taking is a trap, and the behavioral consequences of that understanding, reduced initiative, increased self-protection, accelerated attrition among the most capable people who can find environments where the culture is not a trap, are the most reliable predictors of long-term organizational performance decline.

Steve’s own experience at Spreenklr, learning to manage the behavioral signals he was sending through his urgency, his erratic responses to organizational problems, his public offers of rewards for solving immediate crises, is the practitioner’s version of the same lesson. The team that observes its leader behaving in ways that contradict the stated values is not simply noting a discrepancy. It is recalibrating its understanding of what actually governs organizational decisions, and that recalibration produces behavioral changes in every person who has concluded that the stated values are not the operational ones. The leader whose behavior communicates steadiness, patience, and the willingness to absorb uncertainty without projecting it onto the organization is building the psychological safety that determines whether the workforce’s behavioral signals are honest or managed, and that distinction determines whether any measurement infrastructure deployed in the organization produces accurate data or a carefully curated performance.

The connection to workforce behavioral intelligence is in what leadership behavioral consistency produces as a data environment. An organization where leadership behavior matches leadership rhetoric generates a workforce that trusts the organizational environment enough to exhibit authentic behavioral patterns. An organization where the gap between stated values and operational reality is wide generates a workforce that has learned to manage its visibility, and the behavioral signal that any platform collects from that workforce reflects the management rather than the authentic condition. MustardHub’s stability pattern data is most valuable in organizations where the behavioral signal it collects is authentic, which means its value is directly correlated with the degree to which the leadership environment produces the trust conditions that make honest behavioral expression safe. The platform investment and the leadership development investment are not alternatives. They are prerequisites for each other.

“What you do communicates to your team what behaviors are expected. And so I just started to become really hyper aware of what I said, what I did, when I did, and all that stuff.”

For platform operators, the leadership behavior observation creates a specific customer success requirement that goes beyond the technical deployment of a behavioral intelligence tool. An organization whose leadership behavior systematically undermines the trust conditions required to produce honest behavioral signal will not get the full value of the infrastructure investment regardless of how well the technology works. Building organizational readiness assessment into the deployment process, and helping customers understand the specific leadership behavior changes that create the conditions for accurate signal collection, is the advisory function that separates a behavioral intelligence platform deployment from a technology installation that produces data without producing change.

The practical implication for HR tech operators is that the most valuable customer for behavioral workforce intelligence is not the organization that already has perfect leadership alignment. It is the organization that is actively trying to close the gap between its stated values and its operational reality, that has leaders who are willing to examine and adjust their own behavior, and that is building the trust conditions required for authentic workforce signal to flow. These organizations are not rare. They are, in fact, the majority of the organizations that any serious platform operator’s sales team is talking to at any given time, because the organizations that have already solved this problem are not looking for new solutions.

Rewarding Failure Is Not a Values Statement. It Is the Specific Organizational Behavior That Determines Whether Capable People Stay.

Steve’s closing argument, that organizations have to be real about how they reward failure if they want to keep and attract people who take the risks that drive performance, is the most direct connection in the conversation between cultural rhetoric and workforce outcomes. The gap he described is precise: organizations routinely put values on their walls that include language about innovation, risk-taking, and learning from failure, and routinely fire the people who took risks that did not work out. This is not hypocrisy in the cynical sense. It is the predictable result of a reward and accountability architecture that was designed to optimize for short-term financial metrics and that produces, as a structural consequence, the suppression of the risk-taking that long-term performance requires.

The workforce consequence that Steve identified is the one that most directly produces the attrition patterns organizations spend significant resources trying to reverse: the people most likely to take risks, to drive innovation, and to build the capabilities that produce competitive advantage are also the people with the most options. When the organizational environment communicates, through the actual treatment of people who took risks that did not work, that failure is a career-ending event regardless of the stated values, these are the people who leave first. They have the external options to act on that conclusion, and they will do so. What remains is a workforce that has correctly assessed that the safest organizational behavior is to avoid visible risk, and that behavioral norm produces exactly the performance outcomes that the stated values were supposed to prevent.

The behavioral intelligence implication is that the gap between stated and actual organizational values around risk and failure is directly visible in stability pattern data over time. An organization that punishes failure will show a characteristic pattern in its attrition signals: departures concentrated among the higher-performing, higher-initiative employees who took visible risks, and retention concentrated among the employees whose behavioral patterns reflect risk-avoidance and status quo maintenance. This is not a hypothesis. It is a behavioral pattern that runs continuously through the post-hire period in organizations where the failure culture gap exists, and it is detectable through the continuous signal collection that MustardHub provides before it produces the attrition statistics that everyone agrees, after the fact, were predictable.

“We have to be real about how we reward failure. The people who are failing in corporate don’t have jobs anymore. They’re fired. And so if we can find a way to reward failure with guardrails, with good leadership, I think we’re going to keep more people here.”

For platform operators, the failure culture argument creates a specific product use case: using behavioral signal data not just to identify who is about to leave but to identify why the pattern of who is leaving indicates a cultural gap between stated and operational values. An organization that sees its attrition concentrated among its highest-initiative employees and can connect that pattern to specific organizational events, failed projects that produced personnel changes, risk-taking that was publicly encouraged and privately penalized, is developing the organizational intelligence required to address the root cause rather than the symptom. This is the difference between a retention tool and a workforce intelligence infrastructure, and the organizations that understand the distinction are the ones most likely to make the investment that produces lasting improvement rather than a temporary reduction in attrition rates.

The connection Steve drew between rewarding failure and keeping people, and between not rewarding it and losing the specific people whose capabilities matter most, is the behavioral argument for why culture is not a soft concern for HR but a financial concern for the entire organization. The attrition costs generated by a failure-punishing culture are not distributed randomly across the workforce. They are concentrated precisely in the population whose departure is most expensive and whose replacement is most difficult. Building the infrastructure to make that pattern visible, to connect the behavioral signal running through the post-hire blind spot to the cultural conditions driving it, and to route that intelligence to the leaders who have both the authority and the incentive to address it, is the product investment that turns the observation Steve has been making for two decades into organizational action that can actually change the pattern.

The Alignment That Separates Ecosystems That Grow Talent From Those That Lose It Applies With Equal Force Inside Organizations.

The ecosystem-level observation Steve has drawn from his City Tech Tour research, that the single most important factor separating cities that develop strong talent from those that lose it is alignment on a shared direction, maps with precise structural logic onto the organizational level. A city without a stated vision for what it is trying to become, where every institution pursues its own objectives without connection to a shared North Star, produces the same outcome as an organization where every team pursues its own OKRs without connection to a culture and values architecture that makes the whole greater than the sum of its parts. In both cases, capable people who need a direction worth investing in eventually find one somewhere else.

The specific failure mode Steve identified in the Milwaukee context, elected officials and business leaders who have never communicated a vision for where the region is going in technology and innovation, is structurally identical to the organizational failure mode that most people and culture teams are hired to address: a leadership layer that has articulated values in general terms without the specificity required to make those values operational, a set of programs and initiatives that were designed to serve those values but were never explicitly connected to them, and a workforce that is trying to figure out what success looks like without a clear enough North Star to orient toward. In both cases, the absence of alignment does not produce neutral organizational behavior. It produces the exactly the wrong behaviors, as every individual makes locally rational decisions that are collectively inconsistent with the outcomes the organization needs.

The behavioral intelligence connection is in how alignment failure shows up in continuous signal collection at the individual level. An employee who does not have a clear understanding of how their work connects to organizational direction will show distinct behavioral patterns in the post-hire period: decreased initiative, reduced collaboration across team boundaries, and the kind of quiet disengagement that precedes the departure decision without ever producing a formal complaint or a poor performance review. These patterns are not visible to the manager who is reviewing quarterly OKR completion. They are visible in the continuous behavioral signal running through the post-hire blind spot, and they are the organizational-level alignment failure expressed at the individual level where it is actually experienced and where it is actually addressable.

“From startups to the top, everyone just kind of does stuff that they think will be beneficial because they’re not malicious. Their intention is to make the environment and the community better, but there’s no alignment to a North Star.”

The MustardHub behavioral intelligence layer surfaces the individual-level behavioral consequences of organizational alignment failure in the post-hire period, providing the visibility that makes it possible to connect a pattern of individual disengagement to the organizational conditions driving it rather than treating each departure as an isolated event. An organization that sees a pattern of behavioral disengagement concentrated among a specific cohort of employees, hired during a particular period or into a particular role family, can investigate whether that pattern connects to an alignment gap in the organizational direction that cohort was given, and can address that gap before it produces the full cost of the attrition it would otherwise generate. This is the organizational intelligence function that turns attrition risk data from a warning about what is about to happen into a diagnosis of why it is about to happen and what can be done to change it.

For platform operators, the alignment argument creates a product positioning opportunity that most of the current market has not articulated: behavioral workforce intelligence is not just a retention tool. It is an organizational alignment instrument that connects the individual-level behavioral signal running through the post-hire period to the organizational conditions that are generating it, giving leaders the ability to identify where the gap between stated direction and lived experience is widest and most consequential. The organizations that build this capability, that use behavioral signal data not just to identify who is about to leave but to understand what organizational conditions are driving the departure, will make better decisions about every aspect of their people strategy than those treating attrition as an individual-level problem with individual-level solutions.

What HR Tech Operators Should Do With This

1. Build the Infrastructure That Surfaces Attrition Risk Before the Resignation Conversation

The universal finding from Steve’s field research is that no organization has a reliable mechanism for identifying departure risk before the employee has already decided to leave. Build your product around the intervention window that exists before that decision is made, not the offboarding workflow that follows it. MustardHub’s continuous behavioral signal collection surfaces the stability pattern shifts that precede resignation decisions by weeks or months, giving organizations the ability to act within the window where action can still change the outcome rather than after it has closed.

2. Redesign Candidate Evaluation to Include Behavioral Capability Assessment Alongside Credential Screening

The job description research Steve described reveals a market-wide failure: every hiring process is optimized for credential defensibility rather than capability prediction, and the candidates it systematically excludes are precisely the ones that organizations with the judgment to hire them have found to be their top performers. Build into your platform the capability assessment questions Steve identified, including zero-to-revenue trajectory, team-building from scratch, and idea-to-execution history, as structured evaluation criteria alongside standard credential review. Connect post-hire behavioral outcomes back to pre-hire assessment data through MustardHub’s stability pattern infrastructure to build the feedback loop that improves hiring decision quality over time.

3. Make Leadership Behavioral Consistency a Measurable Input to Your Organizational Health Architecture

The gap between stated values and operational leadership behavior is the most reliable predictor of the workforce trust conditions that determine whether behavioral signal collection produces accurate data. Build your platform’s organizational health assessment to include leadership behavioral consistency metrics, not just employee sentiment scores, and design the signal routing to give leaders feedback about the behavioral patterns their own actions are producing in the workforce. The MustardHub behavioral intelligence layer provides the team-level stability pattern data that makes leadership behavioral impact measurable rather than anecdotal.

4. Connect Attrition Pattern Data to Cultural Condition Diagnosis, Not Just Individual Risk Flags

A platform that surfaces individual attrition risk is useful. A platform that connects the pattern of who is leaving to the organizational conditions driving departures is transformative. Build the analytical layer that identifies whether attrition is concentrated among specific behavioral profiles, hired in specific periods, managed by specific leaders, or associated with specific organizational events, and route that pattern-level diagnosis to the people with the authority to address the root cause. MustardHub provides the continuous behavioral signal that makes this pattern-level analysis possible, turning attrition risk data from a warning into a diagnosis.

5. Design for the Alignment Gap, Not Just the Engagement Gap

The organizational alignment failure Steve identified at the ecosystem level, the absence of a shared North Star that produces individually rational but collectively incoherent behavior, operates with identical structure inside organizations. Build your platform’s diagnostic architecture to surface whether individual behavioral disengagement patterns connect to organizational alignment gaps, and route the organizational-level diagnosis to the leaders who can address it. MustardHub’s infrastructure layer embeds within the systems operators already use, making it possible to connect individual behavioral signal to organizational conditions without requiring a separate analytics platform or a dedicated data science team to interpret the results.


Steve Glynn is the founder of Experience Milwaukee, a media platform covering the people and organizations building Milwaukee’s tech and business future. You can follow his work and commentary on LinkedIn and connect with the Experience Milwaukee community at experiencemilwaukee.com.

If this conversation surfaced questions about the behavioral signal your platform is currently missing, MustardHub provides the infrastructure layer that embeds behavioral workforce intelligence inside the platforms HR tech operators already run. Visit mustardhub.com to learn how we surface attrition risk, financial impact, and stability patterns from the post-hire blind spot no transactional system was designed to measure.

Detect flight risk weeks before the resignation lands.

Employers save $3,000+ per employee by catching preventable exits early.

See who's likely to quit

About the author

Estimate your turnover cost

Live insights on your workforce in two minutes. 5 questions, no credit card.

See my $ risk
Stay in the signal.
Original research and fireside chats on behavioral workforce intelligence, in your inbox.
Subscribe