Primary metric definition
One clearly defined metric tied to the actual business decision the engagement is meant to inform, agreed before work begins.
Measurement
Choosing the right metric before the work starts
The wrong success metric makes good work look like failure and bad work look like success. We agree on the right one first.
Before any campaign, redesign, or optimization program starts, someone has to answer a deceptively simple question: how will we know if this worked. Skip that question and you end up arguing about it after the fact, usually with whoever is unhappy pointing at whichever metric moved in their favor. We treat metric selection as its own deliverable, done before the work begins, not as an afterthought pulled together for a quarterly review.
The core distinction is leading versus lagging indicators. A lagging indicator, like revenue or closed deals, tells you the final outcome, but it tells you late, often weeks or months after the actions that caused it. A leading indicator, like qualified form submissions or demo requests, moves earlier and gives you a chance to react before a quarter is already lost. A good measurement plan uses both: leading indicators to steer week to week, lagging indicators to confirm the steering actually worked.
Guardrail metrics exist to catch the damage a primary metric will not show you. If the goal is to increase form submissions, a guardrail might be lead quality or sales-accepted rate, because it is trivial to increase submission volume by loosening the form and just as easy to flood sales with unqualified leads while looking like a win on paper. Every optimization goal needs at least one guardrail that would catch it quietly going wrong.
And then there are vanity metrics, the numbers that feel good to report and mean almost nothing about the business. Pageviews with no bearing on conversion. Social followers with no bearing on revenue. Time on site that could reflect genuine engagement or a confusing interface people can't escape. We will not build a reporting dashboard around a number just because it trends upward and looks good in a screenshot. If a client wants it reported anyway, we say clearly that it is not a signal of anything and should not drive decisions.
What it is
How we choose the metrics that actually matter
Metric selection ties directly to the business decision the work is meant to support.
We start by asking what decision the engagement is actually meant to inform. A metric that does not connect to a real decision, such as whether to keep spending on a channel or whether a page redesign should be rolled out sitewide, is not worth tracking as a primary success measure regardless of how easy it is to pull.
From there we map the causal chain between the leading indicators available early and the lagging outcome the business ultimately cares about. If a leading indicator has historically not correlated with the lagging outcome, we say so and either find a better leading indicator or accept that the program needs more time before results are legible.
We define guardrails alongside the primary metric, not after a problem shows up. If the primary goal is speed of a form completion flow, the guardrail might be completion rate itself, so a faster form that people abandon more often does not get scored as a win. The guardrail is chosen specifically to catch the most likely way the primary metric could be gamed or could mislead.
Finally we agree on reporting cadence and thresholds for what counts as a meaningful change versus normal variation, so nobody is surprised later by how a result gets judged. This agreement is written down before the work starts, precisely so it cannot be renegotiated after the numbers come in.
Fit
Who this is for, and who it is not for
We would rather say no early than sell a program that cannot work.
Right fit
- You are starting a new campaign or program and want agreement on how success will be judged before it launches.
- Past engagements have ended in disagreement about whether the work succeeded because the metric was never pinned down.
- You suspect a metric currently being reported to leadership is a vanity number that does not reflect real business impact.
Not the right fit
- You want us to select whichever metric will look best in hindsight rather than the one that genuinely reflects the goal.
- You need a single universal metric that works across every campaign and channel with no customization. That does not exist honestly.
- You are not willing to also track a guardrail metric alongside the primary one. Without a guardrail, optimization can quietly cause harm.
Deliverables
What a success metrics engagement includes
A written measurement plan agreed before the work starts.
Leading indicator selection
Early-moving metrics chosen because they have a demonstrated or reasonable causal link to the lagging outcome that matters most.
Lagging outcome mapping
The final business outcome the leading indicators are meant to predict, tracked so the connection can be verified over time.
Guardrail metrics
At least one metric chosen specifically to catch the primary metric being gamed or optimized in a way that quietly causes harm.
Vanity metric flagging
A clear written note on which commonly reported numbers do not correlate with outcomes and should not drive decisions.
Reporting thresholds
Agreed definitions for what counts as a meaningful change versus normal variation, set before results start coming in.
How we run it
How we set the metrics
Done before work starts, not reverse-engineered from whatever numbers moved.
Step 1: Identify the real decision
We start with what decision the work is meant to inform, since a metric disconnected from a decision is not worth tracking as primary.
Step 2: Select leading and lagging pairs
We choose an early-moving indicator and the final outcome it should predict, and check whether that link actually holds in the available data.
Step 3: Set guardrails
We define at least one metric that would catch the primary metric being optimized in a way that causes quiet harm elsewhere.
Step 4: Flag vanity numbers
Commonly reported metrics that do not correlate with outcomes are named explicitly so they do not quietly become the de facto scorecard.
Step 5: Document and agree
The full measurement plan is written down and agreed before work begins, so success is judged against a standard set in advance.
What is the difference between a leading and a lagging indicator?
A leading indicator moves early and predicts a future outcome, like qualified form submissions. A lagging indicator, like closed revenue, confirms the outcome after the fact. Effective measurement uses leading indicators to steer and lagging indicators to verify.
What is a guardrail metric?
A guardrail metric is tracked alongside the primary success metric specifically to catch when optimizing the primary metric causes quiet harm elsewhere, such as lead quality dropping while lead volume rises.
Where this connects
Where metric selection connects
The metrics chosen here shape every other piece of the measurement pillar.
Once metrics are defined, they need to be captured correctly through GA4 setup so the numbers being judged actually reflect reality.
Specific funnel and goal metrics rely on properly built conversion tracking rather than default platform events that miss the real action.
Chosen metrics directly determine which hypotheses matter inside strategy and optimization since a test is only worth running if it moves an agreed metric.
Segment-level findings that inform metric selection often come from analytics insights work done before a program is scoped.
When the primary metric is lead quality rather than volume, it typically ties back to LinkedIn ads targeting decisions further up the funnel.
Questions
Success metrics questions we get asked
Agree on the metric before the work starts
So success is judged by a standard set in advance, not argued about afterward.