Project workflow
Organize project references into evidence for decisions
Connect design examples, technical documentation, competitor research, and investigation notes to a project goal and a clear decision-making criterion.
Separate collecting from deciding
Early in a project, broad collection is useful. Keeping that same undifferentiated list during the decision stage makes comparison difficult. Give newly captured material a research status, then promote only the references that actually influence direction. This separates the pool of possibilities from the smaller set of evidence behind a decision.
In the note, replace reactions such as beautiful or looks good with an observable reason. A mobile navigation is simpler, a pricing comparison is easier to scan, or an error state explains recovery clearly. Concrete observations allow a teammate or your future self to evaluate the reason instead of trusting an unexplained preference.
Keep one decision in each collection
A collection for an entire project quickly mixes design, engineering, research, and operations. Narrow it to a decision that can actually be made, such as landing-page message, authentication approach, or mobile card layout. A smaller decision boundary reduces the number of links that must be compared and keeps the discussion focused.
Include both the project name and the decision topic in the collection title so it remains distinct from similar research elsewhere. Use tags for shared technologies or disciplines. This preserves project context while still letting you retrieve the same kind of reference across multiple collections when a later project faces a related choice.
Record counterexamples and reasons for rejection
A library of only admired examples does not explain why other options were rejected. Keep a small number of counterexamples when accessibility, performance, operational cost, or a requirement made them unsuitable. A short rejection note prevents the same alternative from returning later without the context that ruled it out.
You do not need to preserve every rejected link forever. Keep the examples that establish an important comparison, and remove duplicates or sources whose context has disappeared. The value of a project reference library is not its size. It is how quickly it can restore the evidence needed to review or defend a decision.
Turn the collection into a reference after the decision
Once the direction is chosen, update the notes on selected links with the outcome. Record which part you applied and which part you deliberately left out. The collection then changes from a research inbox into documentation connected to the actual implementation, making it more useful for maintenance and future decisions.
Before publishing a collection, check that private notes and sensitive sources are not exposed. Savepond public collections are designed to show limited link information, but the person sharing them still needs to confirm that each source may be shared and that the original attribution and destination remain clear to visitors.