Thu, 10 Sep 2026

Hi friend,

Most mental health companies wait too long to start building evidence.

By the time an investor asks for outcomes or an enterprise buyer requests peer-reviewed evidence, it’s already too late. Evidence generation takes time, so founders who want to be ready make a plan before anyone asks for one.

But many soon realize that the studies they launched are on timelines that don't match the business goals they’re supposed to support. Or maybe they launched the wrong studies, or didn’t ask the right questions to generate the evidence that was actually needed.

To solve this, you need a science roadmap. A science roadmap connects the evidence you need to the business goals it must support, then works backwards to determine what to do, when to do it, and how to get it done.

In this Hemingway Guide, Jen Huberty PhD walks you through exactly how to build a science roadmap for your mental health organisation.

Let’s get into it.

What a Science Roadmap Does

A science roadmap defines the evidence you need to support business goals and places them in a timely and achievable plan. It maps:

business goals
evidence gaps
scientific activities
sequencing and timelines
ownership
and resources

Together, these turn evidence generation into a plan the business can prioritize and execute.

A strong roadmap starts with your company’s near-term priorities. For an enterprise-focused platform, that might be clinical differentiation, payer adoption, and provider trust. For a consumer platform preparing to sell to employers, it might be cost and ROI evidence, category leadership, and product credibility. These goals must come from the business.

The roadmap then defines what evidence already exists and what’s missing, but also where existing questions or methods were not designed to produce the answers the business ultimately needs. Every gap then gets paired with a specific solution.

The five steps below show how to turn these elements into an executable science roadmap.

How to Build Your Science Roadmap

Step 1: Conduct an evidence needs assessment

Start by clearly defining your business goals. Then assess the evidence you already have to support those goals. Review completed and in-progress studies, existing datasets, white papers, publications, product analytics, customer research, and planned work.

The aim is to find both the gaps and the “gold”: evidence that already exists but has not been analyzed or packaged to answer the right business question.

For example, a company preparing for payer conversations may already collect relevant outcomes but never analyzed them for the population, timeframe, or outcomes the payer cares about. A reanalysis could be more useful, and much faster, than launching a new study.

An independent scientific review can be especially helpful here, because it can surface what internal teams may miss. Internal teams naturally know the data best, but an outside expert can test whether the questions, methods, and findings will withstand external scrutiny. It’s important to surface any gaps between what a team believes its data proves and what a buyer will accept.

By the end of this step, sort your evidence into three buckets:

usable now
usable with stronger analysis or communication
evidence that still needs to be generated

Step 2: Determine the science that must be conducted

For each gap identified in Step 1, ask what your company must show and what questions the science needs to answer. Begin with the decision you need a stakeholder to make, then determine what they need to see before saying yes.

To figure out what evidence will move a specific stakeholder:

Talk to the stakeholders yourself, nothing beats first-hand insight
Talk to the people already in those conversations, pulling in sales and growth teams, or chatting with someone external who has insight into what the stakeholder wants
Read the current reports on what that stakeholder is looking for and the trends in that area

Avoid the mistake of defaulting to whatever evidence type is easiest to produce, without confirming that it will actually inform the stakeholder’s decision. That said, there are often efficient options for getting fast results.

If strengthening provider adoption is the goal and the company lacks evidence about how the solution fits into care delivery, start with the most resource-efficient option. Can existing implementation data be analyzed? Can structured provider feedback be collected as part of work already underway? Or is a new feasibility study needed? Then, pair each gap with a defined scientific activity, the corresponding goal it supports, and the expected business impact.

Next, match each gap to a concrete scientific activity.

Those activities might include:

a retrospective analysis of an existing dataset
structured interviews or a one-time survey
an IRB submission
a feasibility or implementation study
a prospective pre-post outcomes study
a publication and dissemination plan.

Choose the lightest approach that can credibly answer the stakeholder’s question. Existing implementation data may be enough to support a provider conversation. A payer seeking evidence of clinical impact may require something more rigorous.

Efficiency matters, but producing more of the wrong evidence is not progress. A company with several surveys and no longitudinal outcomes will not close its credibility gap with another survey.

For every activity, name the business goal, expected impact, owner, dependencies, and specific deliverable. Those activities become the inputs for the timeline in Step 3.

Become a Hemingway Pro Member

If you want more practical guides for building a mental health business, consider becoming a Hemingway Pro member. You’ll be joining more than 500 founders, executives, clinicians and investors who get access to deep industry insights and tactical guides every week. You’ll also get access to our private Hemingway Slack group and invites to our member-only events.

Join Hemingway Pro

Step 3: Work backwards with realistic timelines

Evidence generation almost always takes longer than founders expect.

Customer discovery moves fastest. Feasibility studies take longer once you account for recruitment, data collection, and analysis. Pre-post outcome studies take longer still, and a paper's path from submission to acceptance adds time on top of that.

For planning purposes, add lead time for study design and preparation, recruitment and enrollment, the data collection period itself, analysis and synthesis, and communication or packaging. Data access and cleaning are common sources of delay and should be planned for before analysis can start. Always add a buffer of 20-30%, in case something takes longer than expected. Tight timelines shouldn’t mean cutting corners when it comes to rigor.

Retrospective analysis is generally faster than prospective work, and structured surveys are faster than pre-post outcome studies, so those are the first places to look for speed. But the right choice still depends on where a company already stands. A company with no pre-post data that has only published survey findings won’t gain the credibility it needs by producing another survey. Sometimes the slower, more rigorous study is the one that actually gets the company where it needs to go.

Step 4: Identify immediate priorities

Now you’ll want to sequence your activities. First, figure out what has to start this quarter to support next year's business goals. Where are the bottlenecks or dependencies? For example, an IRB approval that must be obtained before recruitment can start, or a data-sharing agreement that has to be signed before analysis can begin. What can run in parallel, and what has to happen in sequence?

Remember, you do not need to start everything at once. Rank initiatives using three factors:

business impact
lead time and dependencies
required resources

Then balance foundational work that will take longer with quick wins. You want to be building a portfolio of evidence that will support your business over time. Foundational work includes anything that cannot be rushed later, such as IRB approval, a data-sharing agreement, or prospective data collection. Quick wins use assets already available, such as reanalyzing existing data to answer a new business question.

The common mistake is prioritizing only what feels urgent today. Long-runway initiatives are easy to postpone, but one day they’ll be needed urgently, and it’s often too late.

Step 5: Assign ownership and resources

Every initiative needs one accountable owner, a budget, and an assessment of the competing priorities that could derail it. Ownership may sit with an internal science lead, a cross-functional team member, or an external science team or fractional scientist.

Ideally, the science function should own the roadmap itself, including its creation and ongoing delivery. If that function doesn’t exist internally, an experienced fractional or external science partner can provide the expertise and ownership needed to move the work forward. Business leaders should still help shape the goals and priorities the evidence is built to support.

The best science roadmaps evolve as a company does. Like any strategy, it should be adapted based on what you’ve learned about what you need your evidence to do and any changes in business priorities. They should also stay aligned with the company’s broader product, commercial, and strategic roadmaps so that evidence development supports the work already underway across the business.

Review the roadmap quarterly, or at minimum twice a year, alongside product and commercial planning. A monthly cross-functional check-in can keep active initiatives moving and surface changes early.

Example: We worked with a Series D company that needed exactly this kind of sequencing. That meant a needs assessment, the roadmap built from the findings, and an implementation plan. What started as plans for a handful of peer-reviewed papers and stronger payer relationships grew into a broader design for UX and prospective studies. The roadmap made the work defensible because every activity connected back to a business decision the company had already identified as a priority.

Common Science Roadmapping Challenges and Solutions

“We do not have the resources”

Start small. Modest evidence generated strategically beats a bigger study scrambled together later. Look first at what you can learn from a retrospective analysis of data you already have before assuming you need something new. It’s also important to prioritize. One study tied to an actual business goal beats three that aren’t. In the long run, reactively trying to generate evidence is much more expensive. Rush fees and missed opportunities cost more than good planning ever will.

“Our business timeline keeps changing”

Science roadmapping isn't about predicting the future perfectly. It's about building evidence infrastructure that holds up across multiple possible scenarios, so that whichever business goal becomes most pressing, you're not starting from zero. That means building evidence that can serve multiple purposes.

You should also review your science roadmap quarterly as your business goals shift, and at a minimum twice per year as part of broader business planning. Break larger studies into stages so that each produces something useful on its own, even if the next stage gets delayed or changes. And be clear internally about which evidence is core to the business and which is opportunistic. A changed timeline shouldn't derail the work that matters most.

Finally, there is value in the process itself. You wouldn’t stop developing a product strategy just because your timelines keep changing.

“Science and business do not communicate”

The science roadmap needs to be reviewed in business planning meetings, and integrated into other roadmaps, not just science team meetings. Business development should co-own the evidence timeline. A regular monthly sync helps, as does translating the science into business language from the start, so no one on the business side has to do that translation themselves. For example, building buy-in can start small. Scientists can sit in on other teams' roadmap meetings, or meet individually with their leaders to understand what they're working toward. From there, look for one small way science can support what they're already trying to accomplish, something they can see and use. Demonstrating value through something concrete can build support for the broader roadmap over time.

“We are too early for this”

Even early-stage businesses should have a minimum viable science roadmap. This should be built around your most immediate business goal and may have very little on it, but it beats no plan at all. Simple customer discovery and needs assessments don't require a large budget. Building the roadmap and the habit of regularly assessing and planning for your evidence needs is a useful muscle to develop.

Conclusion

Companies with a science roadmap enter fundraising, payer, partnership, and enterprise sales conversations prepared instead of scrambling. Each well-designed study also creates infrastructure, data, and learning that make the next one easier. When evidence generation gets integrated into business planning, science stops being a cost center and becomes a strategic asset.

That's the throughline across the examples in this series. When we worked with a national virtual mental health provider, we were able to analyze five million therapy visits and publish four papers in three months. The evidence worked at the speed of the business, not the other way around.

Start with your most important business priorities. Define what that stakeholder will need to see to say yes. Audit what you have, identify what is missing, and build a program of evidence generation that gives you what you need to drive your growth.

Key takeaways

Reactive evidence generation costs more and arrives later than founders expect. A science roadmap aligns evidence with the business goals it's actually meant to serve.
Evidence and business teams often operate on separate timelines unless the company intentionally aligns their work. A science roadmap only works if it's reviewed alongside business planning, not filed away in a science team's own meetings.
Assessing existing evidence is best done using company input and independent review. Evidence gaps should be paired with a solution and then prioritized based on impact, timing, and resources.
Work backwards from business goals using realistic lead times. Most meaningful evidence takes longer to generate than founders expect.
Treat the science roadmap as a living document. Review it ideally quarterly, but at minimum twice per year, and let it evolve as business priorities shift.
The companies that plan evidence generation with the same discipline they apply to product development build a competitive advantage that's difficult for slower-moving competitors to close.

As always, many thanks to Jen for sharing her insights and lessons with us. You can read our other Hemingway Guides with Jen below:

And if you want more practical guides like these for building a mental health business, consider becoming a Hemingway Pro member. You’ll be joining more than 500 founders, executives, clinicians and investors who get access to deep industry insights and tactical guides every week. You’ll also get access to our private Hemingway Slack group and invites to our member-only events.