If you are really making way more than PVWatts estimates over time, then it seems like there is a data mismatch somewhere. PVWatts does show a blue rectangle for the area where the weather comes from - is it off/missing for you? (They note that it is not always available.) The only other question would be whether you are in a location where even in that grid the weather is highly variable, though the boxes I see are for very small areas. You can move the box elsewhere by double-clicking to see what happens (turns out my system would perform much better if I use the weather data due west in Nevada, for example.)
The other thing to be aware of is that, in my experience, the PVWatts daily data is not useful for comparison to daily output. Where I am, it appears to model each day based on its historical weather. As such, I don't get a smooth curve but a lot of ups and downs. Over the course of a month, it typically averages out pretty well, but the daily data does not match. We do, however, have more variability in our daily cloud cover than some places.
It may be that finding the maximum daily number in PVWatts could help you get a sense of whether the system is being correctly modeled, but even then I would expect you to outperform the PVWatts estimates. The better thing to do is wait and look at a month of data. If at that point, your output is significantly above PVWatts (and if the top headline in your local news isn't about the unexpected string of cloud-free days) then there may be an issue, either in weather data or in how the data was entered into the PVWatts model.
