Repositories are a systems problem, not a tooling one
Setting up a knowledge repository isn’t about picking the right tool, it’s about designing a system that respects how people create, access, and use knowledge.
For every user insights knowledge management system I have worked on, I found it encompasses at the very least, the repository tool, the insights generation programs (research streams), the content as an asset, and trust and enablement of the insights producers.
Forming an insights repository is about designing a living system that connects people to insights at the speed of their curiosity.
I wrote this playbook for myself to set up a systemic, scalable, and human-centered knowledge repository.
Stage 1: The tools are enablers, not the solution
A good repository tool stores knowledge.
A better tool takes accessibility and reader engagement to a delightful level.
Heuristics for tool selection (aka making a business case for more than a folder drive)
Findability > Storage
“Can people find what they need faster than they can ask someone?”Support insights, not just data
“Does the tool elevate meaning and feelings over data points?”Enables insights producers(contributors) as much as insights consumers(audience)
“How easy is it for someone to add new insights, consistently?”Workflow and systems integration
“Does it meet people where they already work?”Documents and signals use, reuse, and share
“Can we see if insights are being read, referenced, or reused?”Enables shared vocabulary
“Does it help teams converge on consistent terms and references?”Scales quality without big friction
“Is there a way to maintain quality without adding bureaucracy?”
“Will this still work when you have 10x the content or contributors?”Supports psychological safety for sharing
“Does the tool feel like a safe place to share (and label) outputs without a quest for perfection?”Feels inspiring to use on the daily
“Would a curious product manager, marketing decision maker, or designer want to spend time here?”
Tools matter but they’re only part of the puzzle. Ask,
Who are your insights producers and consumers?
Where and how do these people already work?
Stage 2: Connect research systems with intention
Repositories cannot be useful in silos. Typically insights systems are spread over 3 categories,
Input – research outputs, continuous discovery, analytics
Storage – repositories, folders, CMSs
Consumption – Slack and Teams threads, decks, townhalls
Use manual or software integrations as bridges to ensure information flows between these systems without stagnation.
Heuristics to guide connection of research systems
Insight overload prevention
“Does the system help distill, not just accumulate?”Insights convergence
“Can multiple sources point to the same underlying truth?”Relevance
“Can we tell what’s fresh, what’s stale, and what’s evergreen?”Tool interoperability
“Can the systems talk to each other without duct tape?”
Stage 3: Treat knowledge like a product/asset
Treat knowledge like a product: manage it, measure it, and evolve it continuously.
A crucial aspect is to set the right success metrics. Some metrics I have used in the past were:
Heuristics to inform governance and monitoring
Ownership exists at all relevant levels
“Does every part of the system have a clear owner?”Knowledge quality is consistent and high
“Does content meet a baseline standard of usefulness and credibility?”There is a feedback loop
“Can users flag issues, suggest changes, or request new features in the system?”Governance is documented and visible
“Are the rules of the system known and followed?”Usage is measured for decisions, not for vanity
“Do you track how knowledge is accessed, reused, and shared?”
Stage 4: Create a psychologically safe learning culture
If there is one thing I want readers to take away from this post, it is this…
Knowledge flows when people feel safe to share. Insights hide in folders when people feel judged.
To enable a learning and sharing culture:
Provide templates, trainings, guidelines, and frameworks to generate valuable content
Lower the bar for contributions at the beginning and raise the bar with increasing maturity of the repository.
This is a tricky strategy as you are trying to balance amount of insights, engagement, and feedback on a tight rope.Create feedback loops to monitor and raise quality
Have a quality marker system that recognises and accepts ‘early signals’, observation trends, and all the way to ‘high confidence insights’.
Establish a recognition and reward system for insights contributors (frequent contributors, most impactful insights, etc.)
Psychological safety is the invisible force behind insight-rich, learning cultures.
Without it, people won’t share what they’ve learned, won’t question assumptions, or admit what they don’t know.
Next: AI
Almost all tools and individual processes now use AI to speed up time to insights. The next post will touch upon how researchers and PwdRs use AI in their work-flows but what they are missing.
Soma partners with organisations to design systems that help teams make user evidence-led decisions.
If you want an advisor or an interim partner and want to see how I can help, get in touch.
Great overview with good questions to reflect on :)