Your internal database of existing customers and leads is your most receptive audience and should perform the best. Use this group as the ultimate litmus test for any new offer. If it fails to resonate with this warm audience, it is highly unlikely to succeed with colder, external audiences, signaling that you should not invest further.
Before optimizing a poor-performing offer, ask if doubling its performance would make it a success. If a 100% lift still doesn't meet goals, optimization efforts are wasted. It's more effective to discard the offer and create a new one, as incremental tweaks are unlikely to yield more than a 100% improvement.
Teams can become attached to their own ideas, believing an offer is great despite poor performance data. The market, not internal opinion, is the ultimate arbiter of an offer's value. When tests show an offer isn't working—especially with your best audience—it is critical to trust the data and move on, rather than throwing more money at it.