Defining the Right KPIs and KRAs for Your In-House Performance Marketing Team vs a Performance Marketing Agency

Key Takeaways

• People optimize what they are measured on, so the KPIs and KRAs you set are the single biggest lever on what your performance function actually does.
• A KRA is the broad area of ownership; a KPI is the specific measurable number within it — most companies conflate the two and measure badly as a result.
• Do not set the same generic metrics for an in-house team and an agency: their roles, incentives, and accountability are fundamentally different.
• An in-house team's KRAs are broad — the whole engine, measurement, creative, and compounding institutional knowledge — with KPIs including business outcomes and capability-building.
• An agency's KRAs are bounded to contracted scope, with KPIs tied deliberately to your real business outcomes rather than vanity metrics, because its incentives must be aligned through the metrics.
• Always reconcile an agency's reported numbers against your actual revenue and profit, so you reward real incremental results — not platform-inflated ROAS that does not reach your business.

You Get What You Measure — So Measuring Wrong Is Expensive

The most reliable law in managing a performance marketing function is that people optimize what they are measured on. Set a media buyer a target of platform-reported ROAS and they will optimize platform-reported ROAS — including in ways that inflate the number without helping your business, like leaning on retargeting and branded search that would have converted anyway. Set an agency a target of lead volume and they will produce lead volume, including low-quality leads that never become customers. Set an in-house team a target of this month's cost per acquisition and they will hit it, including by starving the long-term capability-building and brand work that does not pay off within the month. In every case, the metric shapes the behavior, and a badly chosen metric shapes bad behavior in skilled, well-meaning people. This is why defining the right KPIs and KRAs is not an administrative task but the single biggest lever you have on what your performance function actually does.

And here is the specific mistake this guide exists to fix: companies set the same generic metrics for an in-house team and an agency, as though the two were interchangeable, when their roles, incentives, and accountability are fundamentally different. An in-house team is part of your company, aligned with it by employment, accountable for the broad, long-term ownership of your paid engine and the capability that compounds inside the business. An agency is a separate business, aligned with you only to the extent your contract and metrics make it so, accountable for a bounded scope, and reporting numbers that must be reconciled against your reality because their interests and yours are not automatically the same. Measuring both with the same scorecard either over-scopes the agency into things it does not own or under-measures the in-house team on the broad ownership that is the whole point of having it.

So this guide does three things. It clarifies the distinction between a KRA and a KPI, which most companies conflate and which you must separate to measure well. It defines the right KRAs and KPIs for an in-house team, whose measurement should reflect broad ownership and long-term capability. And it defines the right KRAs and KPIs for an agency, whose measurement should reflect bounded scope, deliberately aligned incentives, and reconciliation against your real outcomes. Read it before you write a scorecard, because a scorecard that measures the wrong things will drive good teams and good agencies to do the wrong work, and you will not understand why your well-staffed, well-intentioned function is producing numbers that go up while your business does not.

KRA vs KPI: The Distinction Most Companies Get Wrong

Start with the distinction, because it is conflated constantly and the conflation causes muddled measurement. A KRA — Key Result Area — is the broad area of ownership and responsibility that a role or partner is accountable for: the domain, the outcome-space, the thing they are on the hook to deliver in the widest sense. A KPI — Key Performance Indicator — is the specific, measurable number within a KRA that tells you whether that area is actually being delivered. The KRA is 'own profitable customer acquisition through paid channels'; the KPIs within it are the specific measures — incremental customer acquisition cost, contribution margin after ad spend, payback period — that indicate whether the KRA is being met. The KRA is the responsibility; the KPIs are the evidence.

Getting this relationship right matters because each fails without the other. KRAs without KPIs are vague accountability — 'own growth' with no way to tell whether growth is being delivered well, which lets underperformance hide and makes reviews subjective. KPIs without KRAs are metrics detached from ownership — a scorecard of numbers with no clarity about who owns what and why, which drives narrow optimization of the numbers without accountability for the broader area they are supposed to indicate. The right structure is a small number of clear KRAs, each with a small number of KPIs that genuinely indicate whether that area is being delivered, so that the KPIs roll up into meaningful accountability rather than floating free. The table below shows the relationship with examples.

A second reason the distinction matters: KRAs are where the in-house vs agency difference is largest. Two functions can share a KPI — both an in-house team and an agency might be measured on incremental CAC — while having very different KRAs, because the in-house team owns a far broader area (including the measurement foundation, the creative capability, and the compounding knowledge) than the agency, whose KRA is bounded to the scope you contracted. Confusing the two leads companies to either hold an agency accountable for KRAs it does not own or fail to hold an in-house team accountable for the broad ownership that justifies its existence. Keep the levels straight: define the KRAs first, per model, then the KPIs within them.

KRAs and KPIs for an In-House Team: Broad Ownership and Capability

An in-house team's defining characteristic is that it owns broadly and is accountable for the long term, because it is part of the company and aligned with it by employment. So its KRAs should be broad, covering not just this month's acquisition but the whole paid engine, the measurement foundation, the creative capability, and — critically — the compounding institutional knowledge and capability that are the entire strategic reason for building in-house. If you measure an in-house team only on short-term acquisition KPIs, you under-scope it to what an agency does, and you fail to hold it accountable for (and therefore fail to get) the broad ownership and compounding capability that justify the fixed cost of building in-house in the first place. The in-house KRAs should reflect that the team owns the engine and the asset, not just the campaigns.

Within those broad KRAs, the KPIs should include business outcomes and capability-building, not just efficiency metrics. Business-outcome KPIs — profitable, incremental growth measured as contribution margin after ad spend, incremental customer acquisition cost reconciled against finance, and payback period — keep the team honest that the point is the business, not the dashboard. Capability KPIs — the reliability and coverage of the owned measurement, the health of the creative testing system, the reduction of key-person risk through documentation and cross-training, the growth of institutional knowledge — keep the team accountable for the long-term asset it exists to build. Because the team is employed and aligned with the company, you can and should measure it on this longer horizon; you are not trying to align a separate business through a contract, so you can hold it accountable for the compounding value that a purely outcome-based agency metric would miss.

A caution on in-house KPIs: precisely because the team's alignment is structural rather than contractual, the risk is not vanity-metric gaming but short-termism and narrowness if the KPIs are set too tightly on immediate efficiency. An in-house team measured only on hitting a monthly CAC target will, like anyone, optimize that target — potentially at the expense of the creative pipeline, the measurement investment, and the brand and capability work that pay off later. So the in-house scorecard should deliberately balance short-term outcome KPIs with longer-term capability and asset KPIs, reflecting the broad, compounding ownership that is the whole point of in-house. Measure the in-house team on the engine and the asset over time, not just the numbers this month, or you will get the numbers this month at the expense of the asset you built the team to create.

KRAs and KPIs for an Agency: Bounded Scope, Aligned Incentives, Reconciled Outcomes

An agency's defining characteristics are the opposite in the ways that matter for measurement: its scope is bounded to what you contracted, and its incentives are not automatically aligned with yours because it is a separate business. Both facts shape its KRAs and KPIs. The KRAs should be bounded to the scope the agency actually owns — the channels, disciplines, and outcomes you contracted them for — rather than the broad ownership you would give an in-house team. Holding an agency accountable for KRAs it does not own (the whole business's growth, capability-building it was not hired to do, measurement infrastructure you kept in-house) is unfair and unmeasurable; holding it accountable for the scope it did contract for is clear and enforceable. Define the agency's KRAs as precisely the area you hired it to own, no more and no less.

Within that bounded scope, the KPIs must be deliberately aligned with your real business outcomes, because the agency's alignment is not structural — it comes only from the metrics and incentives you set. This is the crucial difference from in-house. An agency measured on vanity metrics — platform-reported ROAS, click or impression volume, lead count without quality — will optimize those, and because it is a separate business, it has every incentive to hit the contracted number in whatever way is easiest, including ways that inflate the metric without helping your business. So the agency's KPIs should be tied to metrics that map to your real outcomes: incremental results over what would have happened anyway, contribution after spend, qualified pipeline or revenue rather than raw leads, and cost measures reconciled against your finance numbers. Aligning the agency's KPIs with your business outcomes is how you make a separate business's success depend on your success.

And because the agency is separate and reports its own numbers, reconciliation is non-negotiable: whatever the agency reports must be regularly checked against your actual business results — your real revenue, new customers, and profit — so that you reward genuine incremental contribution rather than dashboard performance. This reconciliation is the safeguard that owned measurement makes possible, and it is the difference between an agency KPI that means something and one the agency can inflate at will. The table below contrasts the two models across the dimensions that matter, and the pattern is consistent: measure an in-house team on broad ownership and compounding capability over time, and measure an agency on bounded scope, deliberately aligned incentives, and business outcomes reconciled against your reality. If you want an agency relationship whose KPIs are defined around your real, reconciled business outcomes from the start, that is exactly how our team structures its scorecards.

Putting It Together: A Scorecard That Fits the Model

The unifying principle is that the scorecard must fit the model, because the model determines the role, the incentives, and the accountability, and measuring against the wrong model drives the wrong behavior. For an in-house team, fit the scorecard to broad ownership and long-term capability: define broad KRAs that cover the whole engine, the measurement foundation, the creative system, and the compounding institutional knowledge; and set KPIs that balance short-term business outcomes with the capability and asset-building that justify the fixed cost of building in-house. For an agency, fit the scorecard to bounded scope and deliberate alignment: define KRAs precisely at the contracted scope; set KPIs tied to your real business outcomes rather than vanity metrics; and reconcile the agency's reported numbers against your actual results so that you reward incremental contribution, not dashboard performance.

A few cross-cutting rules apply to both. Keep the number of KRAs and KPIs small, because a scorecard with too many metrics dilutes focus and lets the function optimize the easy ones while neglecting the rest. Choose KPIs that are hard to game in ways that do not help your business — incremental and reconciled measures over platform-reported vanity numbers — because whatever you choose, the function will optimize it, so choose measures whose optimization is genuinely good for the business. Revisit the scorecard as the function and the business evolve, because the right metrics at one stage may be wrong at another. And in every case, remember the underlying law: you get what you measure, so the scorecard is not a report card you write after the fact but a steering wheel you set in advance, and setting it well is one of the highest-leverage things you can do to make your performance function produce business results rather than good-looking numbers.

The failure this guide exists to prevent is the well-staffed, well-intentioned performance function — in-house or agency — that produces metrics that go up while the business does not grow, because it was measured on the wrong things and optimized them faithfully. The fix is not more oversight or better people; it is a scorecard that fits the model, distinguishes KRAs from KPIs, holds an in-house team accountable for broad ownership and compounding capability, holds an agency accountable for bounded scope and reconciled business outcomes, and in both cases measures the things whose optimization is genuinely good for the business. Get the scorecard right and your function's incentives point at your success; get it wrong and no amount of talent or effort will keep it from optimizing the wrong number. If you want help designing KPIs and KRAs that fit your model and reconcile to your real outcomes, that is exactly the kind of work our team does with companies structuring their performance function.