Efficiency, Efficacy, and Effectiveness: How to Measure What Really Generates Results
Doing it fast, doing the right thing, and generating lasting impact are three distinct competencies — and most dashboard indicators only measure the first. This guide shows how to separate output, outcome, and impact, which KPIs reveal activity and which reveal results, how to connect OKRs to strategy, and how to review indicators that don't support any decisions.
There is a type of company that works hard, delivers on time, has organized sprints, an engaged team, a dashboard full of green charts — and yet does not grow. The problem is almost never a lack of effort. It is the silent confusion between three words that seem like synonyms but are not: efficiency, efficacy (or effectiveness of the goal), and effectiveness (or impact).
Peter Drucker summarized two of them in a phrase that has lasted half a century: efficiency is doing things right; effectiveness (efficacy) is doing the right things. The third, impact effectiveness, is what settles the bill — it asks whether what was done right and was the right thing to do really caused a durable change in the customer and the business. A startup can be highly efficient building features no one uses, and be efficacious meeting a goal that moves no relevant needle.
This guide walks the complete path: the conceptual distinction between the three terms, the practical separation between output, outcome, and impact, the sorting of activity KPIs versus result KPIs, the connection between OKRs and strategy, the cases where efficiency becomes a trap, an end-to-end metric tree example, and a checklist to retire indicators that support no decisions whatsoever.
💡 The biggest mistake founders make: confusing motion with progress. A team that doubled its deliveries without changing retention, activation, or margin did not get twice as good — it got twice as expensive producing the same result.
What is the difference between efficiency, efficacy, and effectiveness?
The three words answer different questions about the same work. Efficiency asks at what cost; efficacy asks if the goal was achieved; effectiveness asks if it was worth it. They are cumulative: there is no effectiveness without efficacy, and isolated efficiency is just speed in a direction no one validated.
The confusion has a concrete cost. Companies that measure only efficiency end up rewarding those who produce the most artifacts. Companies that measure only efficacy hit quarterly goals and lose market share the following year, because the goal was disconnected from what the customer values.
1. Efficiency — Doing Things Right
Efficiency deals strictly with the relationship between what goes in (resources, hours, budget) and what comes out (deliveries, tickets, deployments). It is the easiest dimension to instrument and the most dangerous to optimize alone: every process can become cheaper right up to the moment it breaks the customer experience.
2. Efficacy — Doing the Right Thing
Efficacy measures the degree of achievement of the stated goal. A delivery is efficacious if it was launched on time and met the agreed scope, regardless of cost. However, it depends entirely on the quality of the goal: a poorly chosen goal produces perfect efficacy towards the wrong destination.
3. Effectiveness — Generating Real and Sustained Impact
Effectiveness is the only of the three dimensions that survives the central provocation of investors and boards: "What actually changed in the business or in the customer's life because of this delivery?" It is the permanent behavioral change that consolidates in retention and the P&L.
| Dimension | Central question | Indicator example | Risk of isolated measurement |
|---|---|---|---|
| Efficiency | Are we using resources well? | Change lead time, cost per lead, hours per delivery | Speed without direction and quality drop |
| Efficacy | Did we achieve the agreed goal? | % of quarterly goals completed, SLA met | Hitting goals irrelevant to the customer |
| Effectiveness | Did the result change the business? | Cohort retention, net expansion revenue, avoided cost | None — but requires time and cohort discipline |
Stop measuring effort and start measuring results
In the Shinier Accelerator you build your startup's metric tree, define OKRs connected to the strategy, and track outcome indicators on a single dashboard, with mentoring to separate what is noise from what is decision-making data.
For founders under construction
Use the platform's OKR, Kanban, and Dashboard tools to link each team delivery to an observable customer outcome.
For companies and software houses
Want an indicator system that supports portfolio, capacity, and margin decisions? Talk to the Shinier team.
How to measure output, outcome, and impact without mixing concepts?
The rule of thumb is to look at who is doing the verb. If the actor is your team, it is output. If the actor is the customer, it is outcome. If the effect appears in the business and persists over time, it is impact. "We launched the new onboarding" is output. "70% of new users complete the first setup in 24 hours" is outcome. "90-day retention rose from 41% to 58% in three consecutive cohorts" is impact.
Mixing the layers is the most common mistake in results meetings. It happens because output is immediate, controllable, and comfortable — the team delivers and the number appears in the same week. Outcome depends on the behavior of third parties and takes weeks. Impact takes quarters and requires separating the effect of seasonality. The closer you get to impact, the less control and more informational value you have.
Business Impact (North Star)
Net recurring revenue, margin per customer, cost avoided, cohort retention, and LTV. This is the layer where transformation consolidates in the P&L.
Outcome: Customer Behavior
Activation, adoption of critical features, and time to value.
Outcome: Friction Reduction
Drop in recurring support contacts and increase in eNPS.
The mandatory counterweight
No metric should walk alone. The DORA set is the best public example of this: deployment frequency and lead time measure speed, while change failure rate and time to restore measure stability. Optimizing one pair degrades the other if the change is artificial — and it is exactly this tension that reveals real system improvement. Apply the same logic outside engineering: sales velocity against portfolio quality, content volume against conversion rate.
Which KPIs show activity and which show results?
Activity KPIs describe what the team did; result KPIs describe what changed. Both are necessary — the problem is presenting the former as if they were the latter. A quick test resolves most cases: if the number can go up without any customer having a better experience, it is an activity KPI.
| Area | Activity KPI | Result KPI |
|---|---|---|
| Product | Features delivered in the quarter | Feature adoption in 30 days and effect on retention |
| Engineering | Story points completed | Lead time, change failure rate, and time to restore |
| Marketing | Posts published and impressions | Qualified leads by channel and acquisition cost per cohort |
| Sales | Meetings held | Conversion rate by stage and new revenue per salesperson |
| Customer Success | Tickets answered | Churn, account expansion, and time to first value |
| People | Trainings conducted | Talent retention and ramp-up time for new hires |
Vanity metrics: how to recognize them
These are numbers that only go up — cumulative signups, historical downloads, followers, visits. They have no denominator, they have no cohort, and they have no owner. Since they cannot drop, they never provoke a decision. Replace cumulatives with rates and slices: signups that activated in the week, visits that generated a qualified lead.
Goodhart's Law in practice
When a measure becomes a target, it ceases to be a good measure. A closed tickets target generates premature closing; a deployment target generates artificial delivery slicing. That is why every speed indicator needs a quality indicator next to it and a quarterly review of induced behavior.
How to connect OKRs to strategy and customer value?
An OKR is not a to-do list with a date. The objective is qualitative, memorable, and linked to a strategic choice; key results are numerical, verifiable, and — this is the point almost everyone gets wrong — they must be outcomes, not deliveries. "Launch the iOS app" is an initiative. "Raise mobile activation from 22% to 45%" is a key result.
The bridge between strategy and OKR is the cause-and-effect logic of the Balanced Scorecard: learning and culture enable better processes, better processes generate a superior value proposition for the customer, and the value proposition sustains the financial result. If you can't draw this chain for an OKR, it is a wish, not a priority.
Structure of a well-written OKR
- Objective: an inspiring and specific sentence, without numbers, that anyone on the team can repeat from memory.
- Key Results: from two to four, with a baseline, target, and deadline, always measuring a change in the customer or business.
- Initiatives: the bets you believe will move the key results — they can fail without the OKR changing.
- Ritual: weekly confidence check-in and a quarterly retrospective with scoring and logged learning.
Mistakes that empty the system
- Turning the entire roadmap into OKRs — without discarding, there is no priority.
- Linking OKRs to bonuses, which induces easy targets and political negotiation.
- Binary key results ("done/not done"), which are disguised tasks.
- Defining them at the start of the quarter and only looking at the end, when there is no time to course-correct.
- Cascading top-down without space for the team to propose the path.
🎯 The key result test: if it can be marked as completed without any customer noticing a difference, rewrite it. It is measuring output.
When is an efficient operation still ineffective?
Whenever efficiency is measured inside one part of the system and the result happens outside of it. It's the classic local optimum: the development team delivers faster, but the bottleneck is in customer validation; marketing generates more leads, but sales lacks the capacity to serve them; support closes tickets in minutes without solving the root cause, and the same customer comes back three times a month.
The same trap appears in macroeconomics: producing more units is not a productivity gain if the value generated per hour does not grow. The OECD measures exactly this relationship, and the lesson applies to any startup — volume is not synonymous with value.
There is also the most painful case: a flawless operation building the wrong thing. No efficiency gain makes up for a product that solves a problem that the market won't pay to solve. That's why effectiveness needs to be verified before scaling, not after.
Symptom: green dashboard, flat revenue
All operational indicators hit the target and the financial result does not move. There is a lack of an outcome indicator linking the operation and the customer.
Symptom: invisible rework
Fast deliveries followed by corrections that are not accounted for as a cost. Measure rework rate alongside velocity.
Symptom: shifted queue
One team speeds up and the queue just moves somewhere else. The only way to improve effectively is to improve the bottleneck of the entire flow.
Metric tree example for a startup
The metric tree connects a north star metric to a few drivers and, below them, to the operational indicators that each team controls. It solves the alignment problem without micromanaging: anyone can point to where their delivery appears at the top. Below is an example for a B2B SaaS startup in the traction stage.
North Star Metric (Impact)
Net Expansion Recurring Revenue by cohort — grows when the base buys more and churn is controlled.
Driver 1: Customer Retention (Outcome)
What sustains the base over time.
- •Product: Critical feature adoption in the first 30 days.
- •CS: Resolution of onboarding blockers (SLA).
Driver 2: Upsell / Expansion (Outcome)
What increases the value of the active account.
- •Product: Seat utilization rate (if priced by user).
- •Sales: Discovery meetings scheduled with the active base.
The dashboard cleanup checklist
If your company tracks more than 15 global metrics, you are not managing by data — you are archiving numbers. Run this checklist on your dashboard every semester and remove any indicator that fails two out of three tests:
The Actionability Test
If this number drops by half tomorrow, do we know exactly which process needs to be changed? If the answer is no, it is an observation, not an indicator.
The Ownership Test
Is there one specific person whose performance evaluation depends on moving this number? If the number belongs to everyone, it belongs to no one.
The Customer Test
If we double this number, does the customer notice an improvement in the experience? If not, it is an efficiency metric that needs an effectiveness counterweight.
"Efficiency is doing things right. Efficacy is doing the right things. Effectiveness is doing the right things, right, and having it matter in the end."
Referências
- DRUCKER, Peter F. The Effective Executive. It is a reference because he coined the distinction that structures this article: efficiency is doing things right, effectiveness is doing the right things. Drucker argues that the productivity of the knowledge worker does not come from accelerating tasks, but from deciding which tasks deserve to exist — which turns prioritization into a measurable managerial competency. View book
- KAPLAN, Robert S.; NORTON, David P. Balanced Scorecard. It is a reference because they proposed measuring performance across four connected perspectives — financial, customer, internal processes, and learning — linked by cause-and-effect relationships. It is the methodological foundation of the metric tree: no operational indicator is valuable alone; it must explain how it contributes to a business result. Balanced Scorecard Overview
- DOERR, John. Measure What Matters. It is a reference because he popularized OKRs as a focus system: few qualitative objectives, numerical and verifiable key results, short cycles, and total transparency. The book insists that a key result is not a to-do list — if the item can be checked off without changing anything for the customer, it is an output disguised as an outcome. What Matters OKR Guide
- DORA. Four Keys Metrics (DevOps Research and Assessment). It is a reference because, with years of research across thousands of organizations, it defined four balanced software delivery metrics — deployment frequency, lead time for changes, change failure rate, and time to restore service. It is the best public example of a set that prevents optimizing speed while destroying stability. Learn about DORA metrics
- OECD. Productivity Indicators. It is a reference because it standardizes how countries and sectors measure productivity — output per hour worked, multifactor productivity — and shows the macro trap that repeats itself in the micro: growing production without growing value generated is not real productivity gain. OECD Productivity Indicators