The last mile: Turning evidence into action

That question was the spark for the very first Wonkcomms Nairobi, co-hosted with Busara last month — a brilliant evening of speakers from Shujaaz, J-PAL Africa and Busara (and a very engaged community!), each coming at the challenge differently but landing on the same conclusion: by the time research is ready to publish, many of the choices that decide whether people can use or act on it have already been made.

Which means waiting to be handed a finished piece to promote is already too late. As communications leaders, we need to be in the room while audiences, outputs, relationships and routes to action are still being shaped — and to judge success by what the work helps people do, not how far it travels.

Start with the person, not the audience

“Audience” is one of those words we use so often it can stop meaning much.

Jessica Tradati, Strategic Communications Lead at Shujaaz, described a more specific approach. Shujaaz starts with what it calls GroundTruth: rather than beginning from audience assumptions, it listens to young people across the country before deciding the who, what or how of a campaign. That’s especially important in a country like Kenya because context differs so much from county to county — under Kenya’s devolved system, counties run their own services from health facilities to primary care, so the decisions and constraints people meet vary by where they live.

Knowing your audience also means understanding the moment they meet you. Data from Shujaaz’s sexual-health chatbot showed most messages arrive between midnight and 5am. Clinics and other providers are closed then, and for a lot of these young people the chatbot is the only place they can talk about this without being judged. Many aren’t looking for a referral, they just want to be heard. So the aim of the chatbot isn’t a single message that lands, it’s keeping that conversation going so they feel able to seek help the next day, or whenever they’re ready.

We can usually tell you how many people read something. We can only sometimes tell you what they did next. Success should be judged by whether communications contributes to the action the organisation wants — a service sign-up, a briefing request, an invitation into a policy discussion, evidence cited in a decision — not by reach alone. None of these prove communications caused the result, but they are better evidence of its contribution than likes, views and downloads, which only show something travelled. AMEC’s Integrated Evaluation Framework helps make that shift, starting from the organisation’s objectives and working back to what communications needs to achieve.

Build the relationship before you need it

Sakshi Hallan, Policy Manager at J-PAL Africa, made the case for evidence generation running through the policy lifecycle, not bolted on once a programme is designed.

J-PAL’s new Evidence Champions Network, launched in May, puts that into practice. It targets middle managers inside government — department heads, deputies, others close enough to the work to see where evidence could be applied, with enough standing to move it through an institution. A well-placed champion interprets evidence for colleagues, keeps it visible through a long process, and connects it to those who will act — mattering more than the most senior name on the org chart because they know both the formal process and the reality behind it: who drafts the paper, who advises the decision-maker, where resistance will come from, who implements etc.

Michael Onsando from Busara came at the question through institutional relationships. Busara’s long-standing relationship with NACOSTI, Kenya’s research regulator, had done more for its work than any single dissemination push; in May 2026 NACOSTI accredited Busara’s Institutional Scientific and Ethics Review Committee to review protocols in the humanities and social sciences — a relationship shaping the conditions for research before there was anything to disseminate.

In the East African settings many of us work in, access and trust are built through repeated contact, so relationships are part of the work itself, not just a line on a dissemination plan. And a relationship cannot sit with one contact: people change jobs, roles shift, institutions reorganise. We need connections across levels and functions, and more than one person who knows the history.

Relationships also shape who should carry the message. As one contributor from the audience shared: a video series persuaded a sector to adopt a practice not through the organisation’s own voice but through fellow practitioners describing how they had adopted it — peer credibility carrying the message through networks the organisation could not reach itself.

If the first time you invest in a relationship is when you need someone to act on your evidence, you are asking the last mile to do far too much. That trust has to be built earlier.

Some of the groundwork sits inside your own organisation

The “last mile” sounds as though the hard part happens outside the organisation — with policymakers, practitioners or the public. Michael shared a palliative-care example that complicated that. The research pointed to an obvious next step, a brochure for hospital caregivers, but what held it up were mundane problems: a budget that hadn’t accounted for printing and a licensing requirement no one had scoped. The evidence and the idea were both there; the organisation still had to carry it through.

Most comms teams will recognise a version of this. The point is not to put comms in every meeting, but to involve it earlier, where evidence is being framed for use. And the responsibility runs both ways: we cannot wait for an invitation and then complain we were brought in too late. We earn our place by helping colleagues understand who they need to reach, testing whether an argument holds outside the research team, spotting practical barriers and asking what should happen after someone meets the evidence.

If we want communications treated as a core function, we also need to connect it to the goals leaders care about — service use, policy influence, funding conversations, partnerships, changes in practice — by agreeing intended outcomes and data access at the start, then showing where communications contributed without claiming sole credit.

By the time someone asks “How do we promote this?”, the most important communications decisions have already been made.

Before you publish: a practical last-mile check

Before signing off a report, campaign or research product, work through these questions so audience, route to action, relationships and follow-through fit together.

• Name the action you want: Complete the sentence: “By [date], [specific person or group] will [observable action].” “Raise awareness” is not enough. And don’t confuse scale with value — a few of the right people taking a meaningful step can matter more than a big reach figure.
• Map who can influence what happens next: Who has final authority, who shapes the recommendation, who does the work, which peers others trust? Build more than one relationship in each institution that matters.
• Choose evidence of progress before launch: Decide what you’ll track beyond reach, and who owns the data: follow-up requests, sign-ups, invitations into discussions, references in decision documents, peer referrals, a documented change in practice.
• Bring communications in while choices are still open: Involve communicators as audience, framing, outputs and route to use are decided — and come with something useful: audience insight, message testing, or a clear view of what could stop people acting.
• Plan and resource the next step: Check budget, permissions, licensing, production, language needs, service capacity and follow-up responsibility. If someone wants to use the evidence tomorrow, the organisation should know who does what.

There may not be one last mile

Taken together, the examples from our first Wonkcomms Nairobi make a single “last mile” feel too neat. Evidence can stall anywhere between research and action: in how the person is understood, through a missing relationship, inside the organisation’s own systems, or where someone tries to take the next step.

So we need to start earlier — name the action we want, then work back to what needs to be true and what could prevent it. That sees communications as a core function long before the research is packaged, while there is still time to build a realistic route from seeing something to acting on it.

A 90-minute conversation was never going to produce a universal fix, but it did give us a better diagnosis!

Wonkcomms Nairobi will keep making space for conversations like this. We plan to meet three or four times a year, and the questions people bring will decide where we go next. Join us!

Sign up to our Wonkcomms mailing list to be the first to know when the next Wonkcomms Nairobi will be!