SaaS Input vs Output Metric Hierarchy

The most common failure mode in SaaS analytics is not a lack of data — it is a lack of clarity about which metrics teams control and which metrics result from their work. When a company sets an OKR around net revenue retention without connecting it to the specific actions product and customer success teams can take, the metric becomes a number that leaders watch helplessly rather than a target that guides daily decisions.

The input/output metric hierarchy is the structural solution to this problem. It creates an explicit chain from the north star at the top to the granular, team-level actions at the bottom, with leading indicators in the middle that give teams early feedback before the business-level output metrics move. Understanding how to build this hierarchy — and how to design OKRs within it — is the difference between a metrics program that drives decisions and one that drives dashboards.

Defining Input and Output Metrics

An output metric is a business result. Net revenue retention, gross revenue retention, customer lifetime value, and annual recurring revenue are all output metrics. They are lagging indicators — they reflect the accumulated effect of decisions made weeks or months earlier. They are also multiply determined: retention is influenced by product quality, pricing, customer success coverage, competitive dynamics, and macroeconomic conditions simultaneously. No single team can own an output metric because no single team controls all the factors that move it.

An input metric is a specific, actionable behavior that a team can influence directly through their work. The percentage of new accounts completing the onboarding checklist within the first seven days is an input metric. The number of accounts that have used three or more integrations in their first 30 days is an input metric. The weekly count of accounts that have reached the activation threshold defined by the product team is an input metric.

The critical distinction is causal direction. Input metrics cause outputs to move (with a time lag and with noise from other factors). Output metrics do not cause inputs to move — they are the result. When companies confuse these directions, they often set OKRs on output metrics and then find themselves watching dashboards without a clear path to improvement.

McKinsey's research on high-performing product organizations found that teams with explicit input metric targets shipped experiments that moved business outcomes 2.3x more frequently than teams with output-only targets. The reason is mechanical: input metrics are close enough to team actions that the feedback loop is short enough to be useful.

The Four-Level Hierarchy

A well-constructed metric hierarchy has four levels, each connected to the next by a causal or predictive relationship.

Level 1: North Star Metric. The single leading indicator that most directly predicts long-term customer retention and revenue. For a detailed treatment of how to select the north star, see the north star metric selection framework. The north star sits at the top of the hierarchy because it represents the company's theory of value delivery: when this metric moves, the business grows sustainably.

Level 2: Output Metrics. Three to five lagging business results that the company tracks at the executive level. Net revenue retention, logo retention, average contract value expansion, and gross margin belong here. These metrics appear in board decks and investor updates. They are real and important, but they are the wrong target for product teams because they move too slowly and are too aggregated. The net revenue retention framework describes how to read output metrics correctly.

Level 3: Leading Indicators. Metrics that move two to six weeks before the output metrics and are predictive of them. Onboarding completion rate, feature adoption rate, and the percentage of accounts reaching activation milestones are typical leading indicators. These are still cross-functional — multiple teams influence them — but they move fast enough to give product and customer success teams a signal before the output metric is affected. OpenView Partners' 2024 PLG benchmark found that companies tracking three or more leading indicators per output metric detected retention risk 6 weeks earlier on average than companies tracking only output metrics.

Level 4: Team-Level Input Metrics. The most granular level — the specific, measurable actions that a single squad can move through their work. These are the right targets for OKRs. A squad working on onboarding might own: "Percentage of new accounts completing steps 1–4 of the setup flow within 48 hours." A squad working on integrations might own: "Number of new integration connections made by accounts in their first 14 days." Each input metric should be measurable from product instrumentation within a week, not a quarter.

OKR Design Patterns That Connect the Hierarchy

The most effective OKR structure for product teams anchors objectives to the north star or output metrics (communicating the "why") and anchors key results to leading indicators and team-level input metrics (defining the "what").

Pattern 1: The Activation OKR. Objective: Increase the fraction of accounts that experience core value within 14 days. Key Result 1: Increase onboarding completion rate from 42% to 65% (leading indicator). Key Result 2: Increase the percentage of accounts making their first integration connection within 7 days from 28% to 45% (team input). Key Result 3: Reduce median time-to-first-value event from 72 hours to 36 hours (team input). This structure gives the squad a clear north-facing objective with measurable, fast-feedback targets. For the activation rate calculations that underpin these targets, see the activation rate guide.

Pattern 2: The Retention OKR. Objective: Reduce early-stage churn by improving product depth among new cohorts. Key Result 1: Increase the percentage of 30-day accounts using three or more features from 31% to 50% (leading indicator). Key Result 2: Increase weekly active usage rate among accounts in months 2–3 from 58% to 72% (leading indicator). Key Result 3: Reduce the share of accounts with zero product logins in any 14-day period from 18% to 9% (team input). The churn connection is explicit but the key results are all actionable by product.

Pattern 3: The Expansion OKR. Objective: Improve the conditions for natural seat and feature expansion. Key Result 1: Increase the percentage of accounts with more than one department active from 22% to 35% (leading indicator). Key Result 2: Increase the weekly share of accounts viewing the billing/upgrade page from 8% to 14% (team input). Key Result 3: Increase in-product upgrade attempts from 45 per week to 90 per week (team input). Each key result connects to a squad action: improving collaboration flows, surfacing upgrade prompts, or redesigning the billing experience.

The Most Common Hierarchy Design Mistakes

Skipping the leading indicator level. Many companies build a two-level hierarchy: north star and team inputs. This is better than no hierarchy, but it misses the middle layer that gives leadership visibility into whether the inputs are working before the north star moves. A leading indicator like activation completion rate is something the CEO can track weekly; individual squad input metrics are too granular for that purpose.

Using correlation to build causal chains without mechanism. A metric that correlates with retention is not necessarily a leading indicator of retention. It may correlate because both are caused by a third factor — for example, accounts with high-touch customer success may complete more onboarding steps AND retain better, but it is the customer success investment that drives both, not the onboarding completion causing the retention. Every link in the hierarchy needs a plausible mechanism, not just a statistical association.

Setting input metrics that require cross-functional dependencies. A squad cannot own an input metric that requires another team to act. "Accounts receiving onboarding emails within 24 hours of signup" is a joint product-marketing metric, not a product squad input. If the email is not sent, the squad's metric suffers regardless of the product work they shipped. Team-level input metrics must be within the squad's direct control.

Failing to calculate the metric hierarchy's break-even. For a leading indicator to be useful, moving it must predict a material improvement in the output metric. If increasing onboarding completion rate from 40% to 65% only improves 90-day retention by 0.5 percentage points, the leading indicator is not connected tightly enough to the output to justify the OKR investment. Before locking in a leading indicator, run the math: "If we move this leading indicator to our target, what improvement in the output metric do we expect, and is that improvement large enough to matter?"

Building the Hierarchy from Scratch

For companies that have not built a metric hierarchy before, the construction process follows a specific sequence.

Start at the output metrics layer. List the three to five business results the company cares most about this year. Prioritize the one that is most at risk — typically retention for early-stage companies or NRR for growth-stage companies. This becomes the output metric the hierarchy is built around.

Work upward to the north star. Ask: what product behavior, if measured in a customer's first 30 to 60 days, most strongly predicts whether they will hit this output metric target? Run the cohort correlation analysis described in the north star metric selection guide. The metric with the strongest retention correlation becomes the north star candidate.

Work downward to leading indicators. Ask: what metrics move four to eight weeks before the output metric and can be read from product data? Candidate leading indicators include feature adoption rates, workflow completion rates, activation milestone counts, and inter-session engagement scores. Validate each with cohort analysis.

Work further downward to team inputs. For each leading indicator, ask each relevant squad: what is the single most important thing your team does that we believe moves this leading indicator? Their answer is the team-level input metric candidate. Validate with experiments where possible — run a test that changes the team's behavior and measure whether the leading indicator responds.

Quarterly Calibration of the Hierarchy

The metric hierarchy is not a permanent document. It should be reviewed quarterly to check three things.

First, are the causal relationships still valid? As the product evolves, the mechanisms connecting inputs to outputs can change. An onboarding flow redesign might invalidate a previous correlation between "days to first login" and retention.

Second, are the team-level input metrics still within the squad's control? Organizational changes, new dependencies, or product architecture shifts can move an input metric out of a squad's direct control without anyone noticing.

Third, are the targets still ambitious but achievable? A leading indicator target that was set when 40% of accounts completed onboarding may need to be revised if the team has already reached 62% — the marginal returns on further improvement may be lower than redirecting effort to a different input. The B2B SaaS KPI dashboard template provides a practical structure for displaying the hierarchy in a format that leadership and squads can both use.

Frequently Asked Questions

Conclusion

The input/output metric hierarchy solves the accountability problem in SaaS analytics: it gives every level of the organization a clear, appropriately-scoped metric to own, and connects those metrics through explicit causal chains. Output metrics belong to the CEO and the board. Leading indicators belong to functional leaders. Team-level input metrics belong to squads.

When OKRs are built on this foundation — with objectives pointing at the north star and key results pointing at leading indicators and team inputs — teams have the feedback they need to course-correct within a quarter rather than discovering at year end that their work did not move the business forward. The hierarchy takes time to validate and requires disciplined calibration, but the alternative — a fragmented collection of dashboard metrics without causal structure — is the reason most SaaS companies track everything and understand nothing.