Digital Health Agency: How to Commission Product and Growth Work

The label ‘digital health agency’ can describe a marketing team, a product studio or both. Commission the capability your outcome actually needs.

Digital health agency commissioning guide with an illustrated system connecting product design, engineering, assurance, brand, growth and measurement, with Pharmacy Mentor logo

Ask five digital health agencies what they do and you may receive five different answers. One shapes brands and demand. Another researches patients and clinicians. A third designs and engineers software. A fourth handles regulated content. A fifth tries to connect the whole service.

That category ambiguity is not a reason to avoid agencies. It is a reason to commission around an outcome, operating model and evidence plan instead of a broad label. The right partner may be one integrated team or several specialists with clear interfaces.

In brief

What should a digital health agency brief contain?

Define the user and business outcome, intended purpose, regulated boundaries, existing technology and data, delivery constraints, approval owners, required capabilities, evidence, ownership, support and exit. Ask every bidder to explain the team and method they would use, not simply present an undifferentiated service list.

  • Separate product, assurance and growth needs before comparing partners.
  • Make clinical, privacy, security and accessibility responsibilities explicit.
  • Retain practical control of accounts, data, code, content and measurement.

Start with the change you need to create

Describe the current situation and what should be observably different for patients, professionals, operations and the organisation. “Build an app” is a solution. “Reduce failed hand-offs between eligibility checking and appointment completion” is an outcome a team can investigate, design and measure.

State what is known, what is assumed and what needs discovery. Include constraints such as procurement dates, integrations, clinical pathways, internal capacity, legacy suppliers and non-negotiable standards. This gives agencies room to challenge the proposed route without making the brief vague.

Map the capability mix

A useful procurement model divides the work into three connected capability groups. Product and engineering covers research, service design, content, UX, architecture, software, integration, testing and operation. Assurance covers intended purpose, clinical safety, privacy, security, accessibility, regulation, evidence and quality management. Brand and growth covers positioning, creative, content, search, media, conversion and lifecycle communication.

Not every commission needs all three at the same depth. A validated product with weak demand may need positioning and growth. A strong audience proposition supported by brittle manual operations may need product discovery and engineering. A new clinical workflow may need assurance leadership before either. Require the lead agency to show where its competence ends and how other specialists will be governed.

Define intended purpose and boundaries early

Document what the service does, for whom, in which context and with what effect. Distinguish information, administration, communication, decision support and clinical functions. Intended purpose influences evidence, risk management and whether software may fall within medical-device regulation; it cannot be retrofitted responsibly after the interface has been designed.

The MHRA's software and AI as a medical device guidance is a starting point for understanding the UK framework. Obtain appropriate regulatory, clinical-safety and legal advice for the actual product. An agency should identify questions and evidence needs, not declare a convenient regulatory answer during a sales call.

Commission research as risk reduction

Ask how the team will learn from intended users, people with access needs, staff performing the backstage work and those who abandon or cannot use the existing route. Research should test the problem, language, trust, workflow and consequences of failure, not merely preferences for screen designs.

Look for a traceable path from evidence to decision. Research findings should change priorities, prototypes, content and acceptance criteria. Pharmacy Mentor's healthcare UX design guide explains how service blueprints, accessible prototypes and operational hand-offs turn user insight into delivery decisions.

Make assurance part of delivery

Do not create a product and then send it through a final compliance gate. Name accountable owners and integrate clinical-safety, information-governance, security, accessibility and regulatory work into discovery, design, engineering and release. Require issues to be recorded, prioritised, resolved or explicitly accepted by an authorised person.

For work intended for NHS organisations, the NHS describes the Digital Technology Assessment Criteria as national baseline requirements covering clinical safety, data protection, technical security, interoperability and usability and accessibility. The applicable route depends on the buyer and product, but the underlying lesson is broader: assurance must produce evidence throughout delivery.

Inspect the proposed team, not the agency logo

Ask to meet the people who will do the work. Review their relevant healthcare experience, roles, availability and decision authority. Clarify which work is subcontracted, where teams are located, how continuity is protected and whether senior specialists remain involved after discovery.

Request examples that resemble the risk and operating model, not just the visual style. A consumer wellness campaign, a patient portal, a professional workflow and prescribing software demand different evidence. Pharmacy Mentor's healthcare software commissioning guide adds detailed questions about architecture, testing and supplier support.

Set a delivery model with decision points

Define the first phase by questions to answer and evidence to produce. A good discovery may deliver a prioritised service blueprint, tested prototype, technical options, risk and dependency log, measurement plan, delivery roadmap and cost range. Agree the decisions that unlock the next phase.

Use short, reviewable increments once delivery begins. Demonstrations should show working journeys and backstage operations, not presentation progress. Acceptance criteria need to cover content, accessibility, security, performance, analytics, error handling, support and data as well as the happy path.

Protect ownership and portability

List every asset: domain, cloud tenancy, repositories, design files, analytics, advertising accounts, consent records, content, data models, documentation, licences, training materials and supplier credentials. State who owns each one, where it is held, who can administer it and how it will be exported.

Prefer organisation-controlled accounts with role-based supplier access. Require readable source, deployment instructions, architecture and data-flow documentation, configuration records, incident history and a supported handover. An exit plan is not distrust; it is basic continuity for a healthcare service.

Separate evidence from theatre

Ask case-study references to explain the starting condition, the agency's exact role, constraints, decisions, delivery evidence, result and what happened after launch. Treat unqualified percentages and award slides cautiously. Results should be attributable enough to inform your risk, not merely decorate a pitch.

For a growth commission, inspect search or campaign assumptions, claim approval, conversion definitions and access to source data. For a product commission, inspect research outputs, quality controls, test evidence, release records and service measures. For an integrated brief, require one measurement model that connects acquisition to safe completion and operational value.

Design the commercial model around uncertainty

A fixed price can suit a bounded, understood deliverable. A time-and-materials model can suit uncertain discovery or iterative product work when priorities and controls are transparent. Retainers can suit continuous optimisation when outputs, capacity, governance and exit are explicit. Do not choose a model because it sounds safest; choose the one that allocates risk to the party best able to manage it.

Compare the total relationship: discovery, build, assurance, licences, hosting, media, maintenance, support, content change, analytics, security testing, future releases and exit. A low initial build price can create expensive dependency if core assets or knowledge remain inaccessible.

Measure one service from discovery to outcome

Define measures at four levels. Experience measures show whether intended users can understand and complete the journey. Safety and quality measures show errors, incidents, accessibility barriers and assurance performance. Operational measures show demand, capacity, hand-offs and staff workload. Business measures show sustainable adoption, service value and cost.

Give every measure a definition, source, owner, review cadence and decision it can influence. Avoid calling a dashboard successful because it reports activity. Pharmacy Mentor's healthcare data analytics guide explains metric contracts, data quality and action ownership.

Use the shortlist to expose assumptions

Give shortlisted agencies the same scenario and invite written questions before the presentation. Score their understanding of the outcome, team, method, assurance, ownership, evidence and commercial clarity. Record the reasons for the decision and the assumptions that must be tested first.

If the brief is mainly demand generation, explore Pharmacy Mentor's healthcare marketing agency approach. If it centres on identity and positioning, see our healthcare branding work. For connected product and growth work, book a consultation with Pharmacy Mentor to shape the commissioning brief before choosing the delivery model.

Frequently asked questions

What does a digital health agency do?

The term can cover product research, UX, software engineering, clinical and technical assurance, branding, content, marketing and growth. Buyers should define the required capabilities and outcome because agencies use the label differently.

How do I choose a digital health agency?

Compare the proposed delivery team, relevant evidence, understanding of users and operations, assurance method, ownership terms, measurement plan and exit route. Use a common scenario and scoring framework rather than choosing on presentation style alone.

Does a digital health agency need regulatory expertise?

The required expertise depends on intended purpose, product and market. The agency should identify boundaries and bring appropriate regulatory, clinical-safety, privacy, security and quality specialists into delivery. Formal determinations need competent advice for the specific service.

Should one agency handle product and marketing?

An integrated team can reduce hand-offs when product experience, service operations and demand are tightly connected. Separate specialists can be stronger when responsibilities, interfaces, decision rights and shared measures are clear. Choose the operating model that best fits the risk and capability need.

Keep exploring

More pharmacy insight

Pharmacy Marketing

Pharmacy Marketing Ideas: 12 Campaigns Built Around Real Capacity

Twelve campaign ideas are useful only when the pharmacy can deliver the service, approve the message and measure what happened after the click.

Read article →
Pharmacy Technology

Pharmacy Chatbots: A UK Guide to Safe Patient Conversations

A chatbot can answer an opening-hours question in seconds. The real test is what happens when the conversation becomes personal, urgent or clinical.

Read article →
Website Design

Healthcare Website Design: A UK Commissioning Guide

A healthcare website project is won or lost before the homepage takes shape: in the service model, evidence, governance and ownership written into the brief.

Read article →