Field notes
Many decisions in a software initiative rest on uncertainty: about performance, about whether a third-party API behaves as documented, about whether an algorithm can be implemented in reasonable time. Deciding before the uncertainty is handled is guessing. Spikes are a deliberate way to buy knowledge: a time-boxed investigation whose only goal is to answer a specific question, so that the main decision can then be made on facts.
Each card has a Purpose, an Allocated time and a Deadline. Keep them small enough to be cheap, and large enough to actually answer the question. Use Status to follow the flow.
Spikes can be used as a question backlog, to keep track of discussions that need to happen. They can describe what must be experimented with before the solution proposal is done, or capture the first experiments and the coming ones.
If there was not enough time to finish a spike the way you intended, that is important information in itself. Write in Result what actually got done before the deadline.
The tab can also be used as a controlled, win-win-based model for deciding on smaller paid experiment assignments. It can be an excellent way to begin a professional relationship.