Suppose two teammates are getting a customer project ready to start. One can answer every question in the meeting and stays late to fix whatever goes wrong. The other notices that information is missing from the order, works with Sales to fix the handoff, and asks a colleague to try the new process until they can run it on their own. By the next meeting, the problem that used to keep someone late has disappeared. A manager who remembers only what happened in the meeting might overlook the work that made the whole team's job easier.
A+ performance means consistently exceeding the expectations of a role while keeping the work sound and making it easier for others to carry forward. I use the term here for a specific role and period. Someone who excels in one job may need practice before taking on another. And someone who has had fewer chances to tackle hard assignments may have less evidence to show.
Look for work that happened outside your line of sight
Start by asking what changed, who contributed, and what evidence you can review. In the kickoff example, I'd look at the handoff record before and after the change, check which meetings slipped because information was missing, and ask the people who took over what they can now do on their own. If Sales fixed the data and the service team tested the new process, give each group credit for what it did.
Motowidlo and Van Scotter distinguished core job performance from behaviors that help people around you do their work, such as coordinating and helping colleagues. In their study of U.S. Air Force mechanics, each dimension added information to overall performance ratings. I apply that distinction by looking at what someone was responsible for delivering alongside what they did to help the team deliver. Motowidlo and Van Scotter (1994)
Collect examples from several work cycles. Ask employees what they did and where they ran into trouble, then ask firsthand observers the same question, such as “What changed in the work because of this person's contribution?” Use the same expectations for people at the same role level and account for the difficulty, resources, and opportunities each person had. When the record is thin, say what else you need to see and give the person a chance to show it. Apply that standard regardless of how many people end up qualifying.
I separate three situations before assigning a grade. Someone who hands off complete work on time is meeting expectations. Someone who does that reliably and fixes a recurring bottleneck so a colleague can take over is showing performance above expectations if improving the process goes beyond what the role calls for. One assignment gives me too little evidence to make the call.
Managers get a closer view of some employees' work than others. Studying manager rotations at one large financial institution, Cullen and Perez-Truglia found that employees with more face-to-face contact with their managers were promoted at higher rates. The setting was one organization. The question I'd borrow for my next review is simple. Who did important work this cycle that I rarely had a chance to see? Cullen and Perez-Truglia (2023)
Hand over the outcome and the authority to deliver it
If the person running a kickoff still needs their manager to approve every scheduling change, the work still queues up with the manager. I'd start with a piece of work whose result is easy to see. Give one person ownership from receipt of a complete, signed order through the customer's confirmation of a kickoff date. Before they start, agree on what they can decide, who can give them the information they need, and which decisions the manager will handle.
In this example, the owner can agree on a date with the customer and prioritize the team's work within the limits you set together. If information is missing, they can go directly to Sales for it. If the customer requests something outside the contract or the kickoff might slip past a date the team promised, the owner brings the manager the impact and the available options. The manager commits to making help available, arranging access to information, and making decisions beyond those limits.
I'd put it this way. “You own the kickoff from the time the complete order comes in until the customer confirms a date. You can agree on a date directly with the customer within the limits we set. If they ask for work outside the contract or we might miss a date we've promised, bring me the options right away. I'll help make that call.”
Bloom, Sadun, and Van Reenen studied how companies in several countries delegated decisions from headquarters to plant managers. Companies headquartered in places with higher levels of trust were more likely to decentralize. The kickoff example is my application at a smaller scale. Give the person the authority the assignment calls for and expand it as you see how they handle exceptions. Bloom, Sadun, and Van Reenen (2012)
Ask how the employee wants to grow. Someone who is good at fixing handoffs may want to redesign a larger process or try managing people. Those paths call for different opportunities and support. Benson, Li, and Shue found that the firms they studied gave substantial weight to sales performance when promoting employees, even though other observable characteristics better predicted how they would perform as managers. Benson, Li, and Shue (2019) When you add a responsibility, agree on which existing work to move off their plate so they have time to own the new assignment.
Talk while there is still time to fix things
When I hand over an assignment, I'd ask the owner to tell me where they'll start and what would prompt them to raise an issue. I'd set a short review cadence at first, perhaps weekly, and space the meetings out once the work is moving reliably. Anything that crosses an agreed boundary should come up right away. If you want people to bring you bad news early, help when they do. Find the person who can make the contract decision, for example, and say when you'll have an answer.
Amy Edmondson studied 51 teams at one manufacturing company and found a relationship between team learning behavior and psychological safety, the shared sense that it's safe to ask questions, admit mistakes, or raise problems. For the handoff, I'd start with “What do we need to know now to keep this commitment?” and let the person finish before jumping to a decision. Edmondson (1999)
When giving feedback, describe what you saw and how it affected the work, then ask what the other person ran into. “On the last two projects, we still didn't have all the customer contacts before the kickoff. The service team had to track them down. What can we fix when the order first comes in?” If the problem starts in Sales, the manager helps fix it there. The owner might agree to flag missing information as soon as they find it. Kluger and DeNisi found that the effects of feedback varied widely and that performance declined in some cases. I'd end the conversation with something both people know they'll try doing differently. Kluger and DeNisi (1996)
Measure how quickly work moves and how often it comes back
If you track only the date the invitation goes out, someone can send it before checking the customer's availability or getting the paperwork in order. In this example I'd measure from receipt of a complete, signed order to the day the customer confirms the kickoff, then check whether the next team actually had what it needed. Track incomplete orders separately so you can see delays customers experience before the clock starts.
| What to track | What to look at | How to use it |
|---|---|---|
| Time from signed order to confirmed kickoff | Median business days, with the number of cases and a list of cases still open | See whether work is moving faster. Separate time waiting on the customer or Sales to find bottlenecks. |
| Quality of the handoff | Number of cases the service team can accept on the first pass and number sent back for fixes, out of the total | Check whether faster scheduling is creating more rework. |
| Whether someone else can take over | Have a teammate follow the process on one real case and record where they needed help | Find steps that still depend on the original owner and credit the people who made the process usable. |
Suppose we give cases in both cycles the same amount of follow-up time. Twenty projects came in during the first cycle. Customers confirmed 18 kickoffs, and the median among confirmed cases was five business days. The next cycle also brought 20 projects, with just 14 confirmed. The median dropped to three days, but six cases were still open and the service team had sent back four of the 14 confirmed cases for rework, up from two of 18. I'd look at the open cases and the rework before calling the process faster.
Before setting a target, check the baseline and the mix of cases. A customer who's ready to start and a customer whose project needs information from several groups are different kinds of work. If the process has just changed and you still don't know where delays come from, set a learning goal. Find where information goes missing and test a fix with a small group of cases. In a complex business simulation, Seijts and colleagues found that a specific learning goal led to better performance than a specific performance goal or a vague “do your best” instruction. In real work I'd set a numerical target once we understood more about the process. Seijts et al. (2004)
At review time, pull up two or three actual assignments and ask who did what, what changed, and which figures reflect the whole team's work. Lead time might have dropped because Sales started sending more complete information, while the owner helped flag exceptions earlier. Give each contribution its due before discussing the next assignment.
To try this with a team this week, I'd pick one assignment that keeps waiting for the manager. I'd tell the owner what they can decide starting tomorrow and put time on the calendar to look at the actual work together. When we meet, we can answer a concrete question. How far has the work moved, and what still has to wait for me?