As Reivers' Uncle Claude1) used to say, “It's better to grab the bull by the tail, not the horns. Then you can let go when you want to.” When a blow is coming through that will provide downwind opportunities, it can be safer to go out after the peak winds have passed and are diminishing rather than early on while the winds are still growing and max roughness of conditions and exact arrival time of peak winds have not yet been determined.
One fundamental aspect of using weather apps and weather websites is to understand that none of them are perfect. With use you can learn to identify local weather patterns for which the apps tend to have greater accuracy and patterns which present problems for them (and consequently should lead you to the clear understanding that such problematic patterns result in less predictable conditions on the water and increased risk for you).
It's not the weather app so much as the model that you choose to use with the app that really matters. In Bellingham, in our experience, the European model (ECMWF) definitely seems to be the most reliable. But these models seem to be most challenged during significantly unsettled weather - particularly in exactly predicting WHEN changes will occur ( or even reach our small zone of usual downwinding if we are just on the edge of some moving system). Know something about which model(s) your app is using (default model, how to change the default, etc.) - we still see long time boating locals who rely too much on the wrong weather prediction models within the apps that they use. And also be aware that unsettled weather is the most unpredictable and that you shouldn't expect complete precision in timing or location from predictions about changing weather.