Choose one practical outcome
Pick an outcome the team can verify, such as a qualified appointment accepted by both parties. Write down what does not count: automatic responses, requests to stop, wrong numbers and messages that need another round of qualification. Keep those categories in the results instead of hiding them.
Record the eligible audience and the number of attempts. Without the denominator, two reply percentages can describe very different experiments.
Keep the comparison reasonably fair
Use the same offer, similar audiences, matching response windows and comparable staff coverage. Record the channel that actually carried the message. A result attributed to an iMessage tool may include SMS fallback.
For a small pilot, treat differences as observations rather than proof of a universal lift. If one group was contacted on a different day or handled by another rep, write that down with the result.
Include the work after the response
Measure how many replies were assigned, how long they waited and whether the next promised action happened. A higher reply count can create a worse customer experience if nobody is available to respond.
Keep a short case log explaining losses and exceptions. Review it with the reps before expanding the budget. The best next step may be improved coverage or clearer copy rather than another line.