Evaluating Team Performance In Engineering

Last month my team was missing deadlines and barely talking during meetings. Today they're crushing targets and leading initiatives on their own. Here's the exact 4-week blueprint I used to turn things around (without firing anyone or using "motivation" speeches): 𝗪𝗲𝗲𝗸 1: 𝗟𝗶𝘀𝘁𝗲𝗻 & 𝗠𝗮𝗽 → Schedule individual coffee chats ↳ No agenda, no laptop, just genuine conversation → Map out each person's frustrations and aspirations ↳ Look for patterns in their feedback 𝗪𝗲𝗲𝗸 2: 𝗥𝗲𝗺𝗼𝘃𝗲 𝗙𝗿𝗶𝗰𝘁𝗶𝗼𝗻 → Eliminate unnecessary meetings → Audit and delete redundant reports → Create clear decision-making frameworks ↳ Who can approve what and when ↳ Where to find key information 𝗪𝗲𝗲𝗸 3: 𝗥𝗲𝘀𝗲𝘁 𝗘𝘅𝗽𝗲𝗰𝘁𝗮𝘁𝗶𝗼𝗻𝘀 → Break down quarterly goals into weekly wins → Give each person direct ownership of outcomes → Create "no-blame" zones for experimentation ↳ Celebrate failed attempts as learning opportunities 𝗪𝗲𝗲𝗸 4: 𝗘𝗻𝗮𝗯𝗹𝗲 & 𝗦𝘁𝗲𝗽 𝗕𝗮𝗰𝗸 → Provide tools and resources they asked for → Stop micromanaging → Let them present their own solutions ↳ Even if different from what you'd choose 𝗧𝗵𝗲 𝗠𝗮𝗴𝗶𝗰 𝗜𝗻𝗴𝗿𝗲𝗱𝗶𝗲𝗻𝘁: Trust them to be adults who want to do good work. 𝗥𝗲𝗮𝗹 𝗥𝗲𝘀𝘂𝗹𝘁𝘀: • Project completion rate up 64% • Team-initiated improvements: 12 in last month • Zero turnover since implementation • Meeting time reduced by 6 hours per week Your team already knows what needs to change. Your job is to clear the path and get out of their way. If this helped you think differently about team performance, share it with a leader who needs to see this.

For decades, engineering teams have been measured by lines of code, commit counts, and PRs merged—but does more code actually mean more productivity? 🚀 Some of the best developers write LESS code, not more. 🚀 The fastest-moving teams focus on outcomes, not just output. 🚀 High commit counts can mean inefficiency, not impact. Recent research from DORA, GitHub, and real-world case studies from IT Revolution debunk the myth that developer activity = developer productivity. Here’s why: 🔹 DORA Research: After studying thousands of engineering teams, DORA (DevOps Research & Assessment) found that the best teams optimize for four key engineering performance metrics: ✅ Deployment Frequency → How often do we ship value to users? ✅ Lead Time for Changes → How fast can an idea go from code to production? ✅ Change Failure Rate → Are we improving quality, or just shipping fast? ✅ MTTR (Mean Time to Restore) → Can we recover quickly when things go wrong? → Notice what’s missing? Not a single metric is based on lines of code, commits, or individual developer output. 🔹 GitHub’s Data: GitHub found that developers working remotely during 2020 pushed more code than ever—but many felt less productive. Why? Longer workdays masked inefficiencies. More commits ≠ meaningful work; some were just fighting bad tooling or slow reviews. Teams that automated workflows (CI/CD, code reviews) merged PRs faster and felt more productive. 🔹 IT Revolution case studies: High-performing engineering orgs measure outcomes, not just outputs. The best teams: Shift from tracking commit counts → to measuring customer value. Use DORA metrics to improve DevOps flow, not micromanage engineers. View engineering productivity as a team effort, not an individual scoreboard. If you want a high-performing engineering org, don’t just push developers to write more code. Instead, ask: ✅ Are we shipping value faster? ✅ Are we reducing friction in our workflows? ✅ Are our developers able to focus on meaningful work? 🚨 The takeaway? Great engineering teams don’t write the most code—they deliver the most impact. 📢 What’s the worst “productivity metric” you’ve ever seen? Drop a comment below 👇 #DeveloperProductivity #SoftwareDevelopment #DORA #GitHub #EngineeringLeadership

I once watched a team's velocity drop by half after hiring their best engineer. On paper, it made no sense. Great business mindset, problem-solving and effective planning. He could debug issues that left others stuck for days. But six months in, the team was shipping slower than before he arrived. The problem was gravity. Every decision had to orbit through him. RFC reviews stalled while waiting for his input. Architectural choices were reopened whenever he disagreed. Junior engineers stopped proposing ideas because they knew he’d find the flaw. He wasn’t trying to slow things down. He was trying to make everything better. And that was exactly the problem. Teams don't scale through individual brilliance. They scale through collective momentum. The engineers who actually multiplied output never looked like rockstars. They wrote boring, obvious code that anyone could maintain. They approved pull requests that were good enough, not perfect. They spent time pairing with people who were stuck instead of cranking out features alone. One senior engineer I worked with had half the commits of everyone else. But every person on his team shipped faster because of him. He asked clarifying questions in planning. He documented the decisions no one else wanted to write down. He unblocked people before they knew they were blocked. His performance review was unremarkable. His impact was everywhere. The best engineers measure their value in outcomes, not output. They care more about whether the team moved forward than whether they personally moved fast. They know that five people shipping with confidence beats one person shipping alone. You don't need a 10x engineer. You need someone who makes ten people better. That's the actual multiplier.

Why is it that engineering teams are so unproductive? Recent discussions of "ghost engineers" and "undifferentiated tasks" offer opinions about the value of where engineers spend their time. These perspectives gain traction because there's a market for them. Executives at all kinds of companies are looking at their engineering spend and wishing they were getting something more from it. Why does it feel like things move slowly and cost so much? These frustrations are real, and executives are paying people to give them explanations and solutions. I don't like the other explanations being sold. But rather than lampooning the competing opinions, today I'm offering my own perspective. Most engineering teams are mismanaged. Rather than trying to root out lazy engineers, executives are better served looking to their leadership for engineering organizations. Engineering time is both costly and [potentially] valuable. Allocating and directing that time towards valuable activities is the essential skill to seeing a return on investment. Here are 5 big ways to mismanage a team, plus straightforward tests of whether you're doing them: 1. Direction—do your engineers know what to do? Test: take a squad of engineers and ask them each to tell you 3 critical team metrics and the 3 most important deliverables this quarter. Compare whether these responses are both aligned with one another and aligned with genuine business value. 2. Enablement—have you set your engineers up for success? Test: if your engineers identify an open-source library that would make their project move faster, how long does it take them to get approval to use it? Your target is "days" or less. 3. Technical standards—are your leaders holding the teams to high performance standards? Test: when a production bug is found, what percentage of the time is there a follow-up that includes AT LEAST one of a) regression tests, b) observability enhancements, or c) changes to team policies? 90%+ is an 'A'. <60% is an F. 4. Automation—are you paying engineers to do things that machines can do more cheaply? Test: survey how long it takes teams to take newly-submitted code and release it to production. What fraction of that time requires active engineer involvement vs. waiting for a tool to finish? Aim for at least a 10:1 ratio of automation : manual interaction. 5. Project planning—are there delivery plans, or just delivery aspirations? Test: look at the delivery plan for a project with a 3+ month timeline. It should be broken down into tasks with a granularity of 2 weeks or less. Tasks should be specific, not "Write code"/"Testing". And progress should be visibly recorded as the project progresses, with completion times grounded in data, not hope. If you're failing these, no wonder you're not getting the results you want from your engineers. A heap of engineers does not an engineering team make. Engineers without direction, empowerment, and support cannot reach their potential.

My Infra team cares about stability. My Feature team cares about speed. If I judge them both by the same “Success Template,” I am failing as a manager. As an Engineering Manager, I oversee multiple teams every day. The standard of work is the same for all of them. The metrics are not, because standardised process ≠ standardised success. Here is how I think about it. Infra teams live in the world of uptime, latency, and throughput. If the platform stays up during traffic spikes, if p95 stays where it should, and if incidents are rare and well handled, they are succeeding. Shipping shiny new things is nice, but it is not the core scoreboard. Feature teams live in the world of users and learning. Their success is adoption, retention, NPS, iteration speed, and the quality of product decisions. If they ship a stable feature that nobody uses, that is not success. Their metric is behaviour change. Platform teams sit in between. Their work shows up in internal developer experience. Build times, onboarding time, test reliability, and how quickly a new engineer can ship something safely. If everyone moves faster and breaks less, they are winning. If I measure all three only by “number of tickets closed” or “story points,” I push them into the wrong behaviour. – Infra starts rushing risky changes. – Feature teams stop talking to users and chase internal OKRs. – Platform teams ship tools nobody adopts. The pilot gets the credit for the smooth landing, but the mechanic keeps the plane from falling out of the sky. If you judge the mechanic by “passenger satisfaction,” you are looking at the wrong data. Good engineering leadership sets a consistent bar for quality and ownership, then provides each team with a scoreboard that accurately reflects the work they are actually doing. If your teams feel like they are working hard and still “losing,” it is worth asking whether the metrics are wrong, not just the people.

I've been in enough performance calibration rooms to recognize this pattern instantly. The spreadsheet says one thing. The business reality says another. Brilliant engineers labeled "needs improvement." High-impact professionals marked "not ready for promotion." Not because they lacked results. Because their work didn't fit the scorecard. Einstein captured it perfectly: Judge a fish by its ability to climb a tree, and you'll call it a failure. Organizations need performance baselines. Absolutely. Without them, reviews become chaotic and inconsistent. But when baselines become your only measurement tool, you systematically undervalue people who create value differently. I've watched this mistake cost organizations millions in lost innovation and retention. One engineer I worked with unblocked integration bottlenecks across three teams. Revenue impact: $2M+ delivered. Her review? "Meets expectations." Why? Her systems thinking was labeled "not technical enough." Her collaborative leadership dismissed as "not assertive enough." The scorecard measured individual code velocity and stand-up dominance. It ignored ecosystem impact. She left six months later. Hired as a principal engineer by a competitor. The fish was redesigning the ocean. The review measured tree-climbing speed. 𝗧𝗵𝗲 𝗠𝘂𝗹𝘁𝗶-𝗟𝗲𝗻𝘀 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 This is how performance should be evaluated - and how you should present your work. → 𝗜𝗺𝗽𝗮𝗰𝘁 𝗟𝗲𝗻𝘀 ↳ What changed because of your work ↳ Revenue protected, risk reduced, time saved, delivery unblocked → 𝗜𝗻𝗳𝗹𝘂𝗲𝗻𝗰𝗲 𝗟𝗲𝗻𝘀 ↳ Who worked faster or made better decisions because you were involved ↳ Team velocity increased, friction removed, alignment restored → 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗟𝗲𝗻𝘀 ↳ Problems you solved differently than others could ↳ New approaches, architectures, or decisions unlocked → 𝗜𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻 𝗟𝗲𝗻𝘀 ↳ What stopped breaking once you connected systems or people ↳ Silos removed, handoffs simplified, value unstuck Before your next performance review: Document your work across all four lenses. Translate strengths into outcomes leadership already values. Stop defending your style - start proving your results. If this resonates and you are a woman in engineering: Join a community turning technical expertise into recognition. https://lnkd.in/gVHn_hmq ❤️ Repost to fix broken performance reviews 🔔 Follow Dima Abu-Khaled for personal and career growth insights

Team capability isn’t binary. It develops through four levels. Most teams get stuck because leaders don’t use a system for progressing people upward. Here is the ladder and how to move someone through it. Level 1: Supervised Can do the work, but not independently. Requires your input before starting, guidance while working, and review before shipping. Typical Behaviors: “Can you check this before I continue?” Waits for approval at each step Unsure of quality standards Many questions, low autonomy Why They Get Stuck: You keep reviewing everything because it feels faster than teaching them to self-correct, or you keep giving them tasks of the same difficulty so they never build judgment. Level 2: Reviewed Executes independently but still needs your review before final delivery. Understands the process but not the quality bar. Typical Behaviors: “I finished this, can you take a look?” Confident in execution, uncertain in evaluation Can produce output but can’t decide if it’s ready Relies on your quality checkpoint Why They Get Stuck: You never defined what “good” looks like, or you keep reviewing out of habit even when their work is consistently fine. Level 3: Independent Executes, evaluates, and ships without you. Knows quality standards and self-corrects. You only step in for novel situations. Typical Behaviors: “I shipped this, here’s what I learned.” Self-evaluation aligns with your evaluation Asks fewer, but sharper questions Confident in both execution and judgment Why They Get Stuck: You don’t expand their scope, or you don’t give them chances to transfer their knowledge to others. Level 4: Teaching Others Performs independently and scales capability by teaching others. Multiplies team performance. Typical Behaviors: “I taught Sarah how to do this.” Documents systems without being asked Runs peer learning or onboarding sessions Reduces your workload by enabling others Why They Stay Here: This is the goal. Maintain and expand their teaching surface area. How to Move Someone Up the Ladder Level 1 → Level 2 Assign the same type of problem three times: First: review every step Second: review at midpoint and end Third: review only at end This develops pattern recognition through repetition with decreasing oversight. Level 2 → Level 3 Define “done” before they start. Show examples of pass vs fail. Then: They self-evaluate before your review When their evaluation matches yours twice in a row, stop reviewing They have internalized your quality bar. Level 3 → Level 4 Have them teach the skill. First document their approach then train a peer Teaching surfaces unconscious knowledge and creates multiplication capacity. The ladder turns supervision into autonomy, autonomy into teaching, and teaching into scale. Year one you are the bottleneck. Year three your Level 4 people are training others while you focus on genuinely new problems. Capability doubles without adding headcount. Build the ladder. Apply the protocol. Capability compounds.

“How do I measure the performance of individual engineers?” Short version, you don’t. Newer engineering managers often ask me this question. It seems like a reasonable question, and it is always asked with sincerity and good intentions. But it comes from a flawed understanding of the nature of knowledge work. We need to move towards team-based performance goals, and away from measuring individual contributors. Tasked with supervising the increase of speed and quality of engineering output, it seems at first that one way to do this would be to identify the so-called “better” engineers and “not so great” engineers. (What one does with the not-so-greats upon finding them is a discussion for another day.) The whole thrust of the question is based on the idea that individual outputs. Modern management should emphasize team collaboration and collective output. In the last century, management principles emerged from the industrial production space, where individuals were often responsible for a certain quantity and quality of outputs for sale (or manufacture for sale). Those days are long gone. Even in manufacturing, Deming argued that performance is largely determined by the configuration of the system, not individuals’ work ethic or performance. Today’s workforce, by and large, “manufactures” knowledge products. The output from knowledge work is designed and assembled collaboratively by skilled cross-functional teams. Nowhere is this more the case than in software engineering. Individual engineers simply do not have much impact on the overall business outcomes achieved by the company. Product development is a team sport, and software products are developed collaboratively. If they aren’t, you’re really just running a feature factory, probably a topic for another post, as well. Traditional management theories by Taylor and Gantt are outdated for knowledge work. Focusing on individual metrics can undermine teamwork and quality. As such, engineering managers should prioritize throughput toward clear business objectives over output by individuals, encourage teamwork and collaboration, and empower teams to improve processes collectively and autonomously. This approach to management requires a mindset shift to one more akin to coaching rather than supervising an assembly line. It's more important to focus on teaching communication and problem-solving skills on the team than on striving for individual performance lifts. If you catch yourself wondering how to evaluate individual engineers as part of a performance improvement effort, think carefully about the "what" and "why" that you’re reaching for.

The metrics we don't measure might be costing our manufacturing operation millions. I recently observed two production teams handling a critical CNC equipment failure. Both had identical technical skills and resources. One resolved the issue within 4 hours with minimal production impact. The other struggled for 3 days, creating a costly ripple effect through their supply chain. The difference wasn't technical expertise. It was psychological safety. When analyzing why high-performing manufacturing teams outpace their peers, we often focus on process improvements, technology adoption, or skills training. But Google's Project Aristotle revealed something operations leaders can't afford to ignore: Psychological safety predicts team performance more reliably than any other factor. Project Aristotle wasn't about feel-good management. It was about creating conditions where teams could apply structured problem-solving methods, learn from failures, and continuously improve without fear. What does this look like on our factory floors? In psychologically unsafe environments, ❌ Engineers hide small issues until they become catastrophic failures ❌ Data gets manipulated to meet expectations rather than reflect reality ❌ Knowledge stays siloed ❌ Structured experimentation never happens because failure isn't an option In psychologically safe environments: ☑️ Problems surface earlier, saving an average of 23% in maintenance costs ☑️ Cross-functional collaboration increases by 35% ☑️ Implementation time for process improvements drops by 41% ☑️ Structured experimentation thrives, making innovation initiatives 67% more likely to succeed Here's 3 three specific practices to try for performance: 1. Separating learning reviews from performance evaluations 2. Establishing clear decision boundaries for experimentation 3. Teaching leaders to respond to problems with curiosity instead of judgment Psychological safety isn't separate from operational excellence. Psychological safety is the foundation that makes it possible. What signals would indicate psychological safety might be our hidden performance multiplier? Here's one simple diagnostic: How do team members respond when equipment fails or metrics slip? With blame and defensiveness, or with collaborative problem-solving? The answer might reveal our next competitive advantage. What psychological safety practices have you seen drive measurable results in your organization? #PsychologicalSafety #Leadership #PersonalGrowth

We’ve all seen the dashboards. Cycle time. Deployment frequency. Incident count. Code churn. DORA metrics if your org is particularly mature. These metrics can be useful up to a point. They show how fast your team is moving, how often they deploy, and how stable things are, but they often miss the most important part of engineering effectiveness: Are we building the right things in the right way, for the right reasons? Here’s what those dashboards rarely capture: ⚠️ How much time engineers spend in decision limbo because priorities aren’t clear. ⚠️ The trust (or lack of it) between engineering and product. ⚠️ The quality of technical debates and whether senior ICs feel empowered to raise concerns. ⚠️ How often work is interrupted, reprioritized, or escalated by leaders chasing quick wins. When I’ve worked with organizations trying to scale, the loudest signal of a healthy team wasn’t velocity, it was clarity. When engineers understand why they’re building something, when they can trace their work to a real customer or business problem, performance improves. Not just in speed, but in outcomes. Engineering metrics matter. But treating them as a proxy for alignment, autonomy, or impact is a mistake. Sometimes, your most "productive" team is actually just the best at working around the dysfunction. Leaders at every level, especially those managing large organizations, need to complement dashboards with real conversations. Spend time with teams. Ask where they're getting stuck. Listen for the gaps that data won’t show. That’s where the real leverage lives. What strategies have you used to build clarity within your teams that didn't involve a stacked bar chart?