Skip forward into hiring notes
RosterAcorn

Quality of hire: 4 weights for your score formula

Set four weights that add up to the full total, and define each input before anyone is scored. Lever's sample formula uses performance at 40%, ramp-up at 30%, hiring manager score at 20%, and retention at 10%, which gives small teams a concrete reference point to adjust from.

Sharp takeaways

  • Start with three to five inputs you can define in writing before you score anyone.
  • Use Lever's 40/30/20/10 split as a reference, then check each weight against your own roles.
  • Score technical and customer-facing roles on windows that match how fast each one ramps.
  • Treat the 80/20 rule as a rule of thumb, not a formula you can apply to your scores.

The four-part weighted score at a glance

Lever's guide to calculating quality of hire builds the score as a weighted sum. Each component score is multiplied by its weight, and the results are added together.

The four components cover different parts of a hire's first stretch of work. The performance score reflects how well the person does the job. The ramp-up score reflects how quickly they reach a working level. The hiring manager score captures the manager's rating, and the retention score records whether the person stays.

Here is how the arithmetic works, without sample figures. Each component score is multiplied by its weight, and the four products are added to give one total. The weights in the guide's formula are fixed shares, so changing any one of them shifts how much that component pulls the total up or down.

The method only works if every component is on the same scale. If one score runs on a ten-point scale and another on a five-point scale, the weights stop meaning what you intend. Convert each input to a common scale before you multiply, and write that scale down so every reviewer uses it the same way. Scores on different scales must be converted to a common scale before they are combined in the weighted sum.

Why three to five inputs beat ten

The guide makes a direct point about the number of inputs: "For many teams, three to five well-defined inputs provide more useful information than 10 inputs that people interpret differently." The reason more inputs can hurt is not arithmetic. It is interpretation. Each additional input is another place where two reviewers can understand the same measure differently.

For a small staffing team, this matters more than it might seem. A long list tends to produce a score that looks precise but reflects a mix of different judgments.

The table below shows where each input count sits relative to the guide's recommendation.

Number of inputsPosition in the guide What happens in scoring
3Low end of the three-to-five rangeEach input is easy to define in writing
5 High end of the three-to-five rangeStill a short list each reviewer can apply the same way
10Above the range, with inputs people interpret differently Reviewers can score the same hire differently

If your current formula has more than five inputs, the practical step is to merge or drop the ones that different people describe in different ways, then write a one-line definition for each survivor.

Comparing the four weights side by side

The guide's example gives each of the four components a fixed share. Compared row by row, the weights rank the components by how much they count toward the final score.

InputWeightWhat it measures
Performance score40%How well the hire does the job
Ramp-up score30%How quickly the hire reaches a working level
Hiring manager score20%The manager's rating of the hire
Retention score10%Whether the hire stays

Performance and ramp-up together make up 70% of the example score. Retention carries the smallest share at 10%. That ordering reflects a choice the formula's author made, and your team should make its own choice based on what matters in your roles.

When you adjust the weights, keep the total at the full whole. Adjust one weight and rebalance the others so the total stays at the full whole, then rerun the same set of hires to compare results. A formula whose weights add up to more or less than the full whole will produce scores that are inflated or deflated, and the error is hard to spot once the numbers are in a spreadsheet. Check the total before you publish any results.

Why ramp speed decides which score you use

The phrase that matters is "ramp at different speeds." If two roles reach full productivity on different timelines, a single time window treats them unfairly.

Consider what happens when you average them. A technical hire who is still getting up to speed at six months and a customer-facing hire evaluated at 90 days are measured against different expectations. Blending the two produces a number that describes neither role well.

The guide warns against blending a six-month score for technical roles with a 90-day score for customer-facing roles when those jobs ramp at different speeds.

The fix is to decide your ramp speed for each role family before you score anyone. Group roles by how quickly they become productive, then assign each group a window that matches its ramp. Once the windows are set, every hire in a group is scored on the same timeline, and comparisons within the group become meaningful.

This also clarifies the ramp-up component itself. The ramp-up score in the formula above only works if you know what "ramped up" means for that role, and that definition depends on the role's ramp speed.

Choosing a 90-day or one-year window

The guide also separates the windows by the question each one answers: "A 90-day score and a one-year score answer different questions." A 90-day score tells you how the hire is doing early on. A one-year score tells you something about the longer run. Neither is more correct in general. They are different measurements.

The table below sets the three windows the guide names side by side.

Score windowRole the guide pairs it withKey point
Six-monthTechnical roles Should not be casually blended with a 90-day score if the jobs ramp at different speeds
90-dayCustomer-facing roles Answers a different question than a one-year score
One-yearNot paired with a role in the guideAnswers a different question than a 90-day score

The practical rule follows from the table. Decide which question each score is meant to answer, then report the windows separately instead of folding them into one number. A team that reports a 90-day score and a one-year score side by side can see whether early results held up over time, which a single blended figure hides.

When you set up the windows, write down the question each one answers next to its length. Anyone reading the report later will then know whether a number reflects early ramp or sustained performance.

Why the 80/20 rule is only a rule of thumb

Recruiting conversations often reach for the 80/20 rule as a shortcut for weighting. The guide is blunt about its limits: "The 80/20 rule in recruiting is a rule of thumb, not a quality-of-hire formula." A rule of thumb can point you toward where to look. It does not tell you how much each input should count.

That distinction matters when you are setting your weights. If you borrow a split because it sounds familiar, you have skipped the step that makes the score useful, which is deciding what each input is supposed to measure for your roles. The 40/30/20/10 split in the guide's formula is a worked example built from defined components, which is a different thing from a general heuristic.

Treat the 80/20 idea as a prompt to question your current weights, not as the weights themselves. Ask whether the largest share of your score goes to the input that matters most for your roles, and whether the smallest share goes to the input you trust least. Then set each weight from your own definitions and check the total against the full whole.

Scoring manager input with a shared rubric

The hiring manager score carries 20% of the example formula, so its quality matters. If one manager rates on instinct and another rates against clear criteria, the input reflects who did the rating as much as how the hire performed.

Build the manager's score the same way. Write three to five rating criteria tied to the work the role actually involves, define what each level on your scale means, and give every manager the same scale.

Keep the rubric short. The inputs advice applies here too: a few well-defined criteria work better than many loosely defined ones. Ask managers to score each criterion rather than give an overall impression, and require a brief note for any score at the extremes so the number can be explained later. A reviewer who reads the note months afterward should be able to tell why the score was high or low without asking the manager to reconstruct it.

Using time-to-fill as context, not as a score

Time-to-fill measures something different from quality of hire. Indeed defines it this way: "In recruiting, time-to-fill is the average time it takes to fill a job opening within your organization." That is a measure of how long an opening stayed open. It says nothing about how well the person hired performs once they start, so it cannot stand in for the performance, ramp-up, manager, or retention components.

Because it describes the hiring process, time-to-fill belongs beside your quality score rather than inside it. A slow fill can reflect a scarce skill set or a demanding role. The number alone cannot tell you which, so reading it alone invites the wrong conclusion.

Keep it out of the weighted sum for three reasons. First, the formula's four components all describe a hire after the start date. Second, adding time-to-fill would give the score a second axis with no defined relationship to the others. Third, the example weights were built around four components, so adding a fifth input would require you to re-derive every weight. Report time-to-fill on its own line in the same dashboard, labeled as process context.

Which one should you pick? Scoring options compared

Three approaches come up when small teams set up a post-hire measure. The table compares them using only the facts the sources give.

OptionFigures from the sourcesBest used when Main risk
Separate 90-day and one-year windowsA 90-day score and a one-year score reported separately You want to see early ramp and longer-run resultsTwo numbers take more explaining than one
80/20 rule of thumbAn 80/20 split You want a quick sense of where to focusThe guide says it is not a quality-of-hire formula

For most small teams, the best starting point is the four-input weighted formula, with the 90-day and one-year windows reported as separate lines. The single score gives you a summary, and the separate windows keep that summary from hiding the difference between early ramp and longer-run results. Lever's guide supports both halves: it recommends three to five inputs, and it notes that a 90-day score and a one-year score answer different questions. Use the guide's formula as your starting weights, then adjust them to your roles.

Avoid choosing the 80/20 split as your main method. It can help you decide where to look, but it does not tell you how much each input should count for your hires. If your team finds itself reaching for it, that is a signal to write down your definitions first.

Reading time-to-fill against your hires

Once time-to-fill is reported separately, read it against the hires it produced. Indeed's definition is an average across openings, so it smooths over individual searches. A team with a few long searches can see its average rise even when most hires are on pace, so look at the individual openings alongside the average.

Pair each time-to-fill figure with its role family. Lever's guide notes that technical and customer-facing roles can ramp at different speeds, and a single average across those two groups mixes two different hiring problems. Keep the averages separate by role family, the same way you separate ramp windows.

Then compare the process figure with the scores. A long time-to-fill followed by strong 90-day scores suggests a patient search that paid off. Both readings are hypotheses to investigate, not conclusions the numbers prove on their own.

Keep the review simple. Each quarter, list the open roles, their time-to-fill, and the scores of hires made from those searches as their windows close. Look for patterns across two or three cycles before changing anything, since one unusual search can distort a single review.

Keeping hire records accurate before you score

A score is only as good as the records behind it. Before any hire is scored, confirm the start date, the role family, the window that applies, and the scale of each component. Those four items decide which timeline and which weights apply. If the start date is wrong, a 90-day score will be measured from the wrong day. If the role family is wrong, the hire may be judged on a window that does not match how that job ramps.

Store the records in one place with the same fields for every hire. Because the weights only work when every component sits on a common scale, check the scale when the data goes in rather than during the final calculation.

Set a reminder for each window's close date when the hire starts. A one-year score cannot be calculated before a year has passed, so the record should show "not yet due" instead of a blank that someone fills in later from memory. Late entries drift toward the reviewer's impression, which is the kind of guesswork a shared rubric is meant to remove.

Limit the records to what the formula uses. Extra fields invite extra interpretation and make the data harder to keep consistent. Lever's guide makes the same point about inputs: a few well-defined ones provide more useful information than many that people interpret differently, so resist adding fields just because they are easy to collect.

For a rubric you can hand to managers, see our Interview Scorecard Example for Consistent Hiring Teams.

Worked example: rebalancing weights for one role family

A small team starts with Lever's split: performance 40%, ramp-up 30%, hiring manager 20%, retention 10%. The team decides retention matters more for its customer-facing roles, so it raises retention to 20%. To keep the total at the full whole, it lowers performance to 30%. The new row reads 30/30/20/20, which still adds up. The team then reruns the same set of hires on both splits and compares the two columns of scores side by side before deciding which to keep.

Related guides: 3 availability reasons to collect at temp hire intake · 3 purposes for each pre-hire background check you order · California hire record fields: 4 items in the state order.

Questions readers ask

Do I need more than five inputs?

Not usually. For many teams, three to five well-defined inputs provide more useful information than 10 inputs that people interpret differently. Extra inputs mainly add places where reviewers can disagree.

What weights should I start with?

Lever's sample uses 40% for performance, 30% for ramp-up, 20% for the hiring manager score, and 10% for retention. Treat it as a reference point and adjust it to your roles, keeping the total at the full whole.

Is the 80/20 rule a formula for quality of hire?

No. The guide calls it a rule of thumb, not a quality-of-hire formula. It can prompt you to question your weights but does not tell you how much each input should count.

Should time-to-fill count toward the score?

Keep it separate. Indeed defines time-to-fill as the average time it takes to fill a job opening, which measures the hiring process, not what the hire does after the start date.

Does a 90-day score tell me the same thing as a one-year score?

No. The guide says a 90-day score and a one-year score answer different questions, so report them as separate lines rather than folding them into one number.

One step you can take today

Write a one-line definition for each input in your current formula, confirm the four weights add up to the full whole, and log the start date, role family, window, and scale for each open hire.

Sources