BLOG POST
Campaign measurement works best when brand teams align objectives, data and methodology from the start.
Measurement has become one of the most high-stakes and intimidating conversations in pharma marketing.
I regularly speak with brand teams who feel like analytics isn’t really their domain. The terminology is specialized, and it’s not always clear how to effectively engage or ask the right questions. When the person across the table has a PhD in statistics and, in some cases, significant authority over media investment decisions, it can be difficult to know how—or when—to push back.
That dynamic carries a real risk. When brand teams aren’t part of measurement design, their media investments can end up being evaluated against the wrong goals and inputs and with the wrong framework altogether. Once a model is built, it can be very difficult to undo the assumptions baked into it.
The good news is that you don’t need to have an advanced math degree to protect your investment. You just need clarity on your objectives and context on which questions are most critical to ask.
Below are eight questions every brand team should raise early.
1 | What is the goal of this campaign, and is that what we’ll actually measure?
This sounds obvious, yet it’s where many measurement strategies begin to drift.
Due to institutionalized measurement approaches, campaigns are often judged against metrics that matter to the business’s bottom line—like ROI—rather than metrics that reflect the original objective. An awareness campaign for a new launch brand might be evaluated on near-term patient starts rather than increased brand awareness. A digital campaign designed to drive qualified traffic might be assessed based on general exposure reach and frequency.
Before any measurement begins, step back and clarify the throughline:
- What specific business objective was this tactic built to support?
- Is the chosen methodology capable of measuring that outcome?
- Do the KPIs directly ladder up to the objective?
When those pieces are misaligned, expectations and performance narratives quickly diverge.
2 | Are we set up to capture the right data?
Even the most sophisticated model cannot overcome flawed inputs. “Garbage in, garbage out” is a common phrase in analytics for a reason. If the data going into the model is incomplete, inconsistent or overly simplified, the output will reflect that.
In life sciences, this often shows up in how exposure is recorded. Many models rely on physician- or office-level “yes/no” flags to indicate that a tactic was present, but that doesn’t tell you how many real patients were impacted.
Ask your analytics team:
- Are we capturing patient-level exposure where possible?
- Are we measuring volume, not just presence?
- Are differences in vendor inputs accounted for?
- Are costs linked to real activity volume rather than evenly proportioned?
If the answer is unclear, the model’s conclusions may be too.
3 | Are we set up to measure the behavior we want to influence?
Sometimes the measurement breakdown happens long before analytics teams get involved.
I’ve seen campaigns where driving website traffic was a core objective, but no UTM parameters were applied during setup. In other words, the campaign launched without the ability to measure whether it achieved its stated goal. That’s not a modeling issue, but a foundational oversight.
Before launching a campaign, it’s important to confirm:
- What specific action are we trying to influence?
- Does my partner team know what specific actions we are trying to influence?
- Is the tracking infrastructure in place to measure it?
- Has that measurement been validated?
Measurement should never be an afterthought.
4 | Does the model reflect clinical context and audience qualification?
Most marketing mix models prioritize reach, frequency and recency. That works well for broad media. It does not work as well for highly personalized, clinically relevant patient engagement, especially in point-of-care environments where DTC and HCP tactics may overlap.
A clinically qualified patient interaction inside a healthcare setting should not be treated the same as a general connected TV exposure.
Ask:
- Does the model account for varying levels of clinical qualification?
- Are highly targeted patient interactions weighted differently than broad exposure?
- Are contextual factors incorporated?
If everything is treated equally, high-intent tactics may be undervalued.
5 | Is measurable patient engagement meaningfully incorporated?
Some channels allow measurable engagement, while others do not.
If a tactic enables confirmed patient interaction (e.g., content engagement, deeper exploration, active selection) it should not be evaluated the same way as passive exposure.
Brand teams should ask:
- Is patient engagement incorporated into the evaluation?
- Are we optimizing for meaningful interaction, not just impressions?
- Does the model distinguish between engagement and presence?
When engagement data exists, it should meaningfully inform performance conclusions.
6 | Are fundamentally different tactics being grouped together?
HCP-facing tactics and DTC campaigns are not interchangeable. They rely on different inputs, serve different objectives and influence behavior differently.
Yet, in some models, they are aggregated together. Similarly, smaller, targeted tactics can be drowned out in models that reward scale and frequency.
It’s critical to ask:
- What channels are being grouped together?
- Do they share comparable goals and data inputs?
- Does the model inherently favor large-scale tactics?
If the frameworks blurs importance differences, the output may skew budget decisions, leading to misallocated investment or valuable, effective media being cut because its impact is understated.
7 | Will this analysis actually inform the decisions we need to make?
Not every measurement tool answers every question.
Marketing mix models are powerful for budget allocation. They are less effective for optimizing creative, refining messaging nuance or evaluating highly targeted investments.
Before you rely on output, ask:
- What decision is this analysis meant to inform?
- Can this tool provide that level of insight?
- Do we need complementary measurement approaches?
If the answer is “the model won’t tell you that,” you may need additional methodologies alongside the model.
8 | Will results arrive in time to be actionable?
One of the most concerning patterns I’ve seen is brands delaying marketing activity while waiting for a model to return results. In one case, nearly all investment was paused for months while teams waited for a marketing mix model output. Teams sat idle, but competitors did not.
Measurement should enable confident action, not prevent it.
Ask:
- When will results be ready?
- Can we act on them in the current planning cycle?
- Are we balancing rigor with speed?
Perfect data delivered too late is often more costly than directional data delivered on time.
Better measurement starts with better questions
Everyone is equipped to ask smart questions about measurement. When brand teams engage early, align goals clearly and challenge assumptions constructively, measurement becomes less intimidating and far more strategic.
The future of pharma marketing won’t belong to the teams with the most complex models—it will belong to the teams asking the clearest questions and measuring the things that matter.
Learn more about how Phreesia Network Solutions can help your team measure what matters.
Karinne Smolenyak is Associate Director of Insights and Analytics at Phreesia Network Solutions. She specializes in applying analytics in ways that are pragmatic, human-centered and grounded in real-world context.
