But do think it I need to delve deeper into this, using real numbers
I'd previously dismissed Agile export for exactly the reasons that b0rn2sk8 mentions, I export a huge amount over the summer months, I think the most I generated in one day was 85 kWh, but can export more due to dumping batteries in the evening
I've looked at this a few times, but cant beat the flat 12p export. The problem as you mention, is that you tend to export most when the price is low on Agile. This can be avoided a bit by either not filling the battery overnight and letting it charge in the day when the export rate is low, or force exporting early on and doing the same. The issue is that with "only" 27kWh, you can have days in summer where you're sending out 50kWh for next to nothing before the evening 4-7pm window. The flat rate wins out then
I finally ran the numbers, downloaded my data for the last 12 months, and also the outgoing Agile rates. Bit of a faff doing it in a spreadsheet, but got there in the end. With my current export pattern Agile export was slightly better than the flat rate, about 1% better over the last 12 months.
To double check I used Octopus Compare, once I'd persuaded it redownload my data, as the first time half my export was missing it came to the same conclusion, in fact both times it was the same conclusion, about 1% better, but for year to date it's almost 3% better, likely due to reduced export in November and December, which will pull it back down.
I'm sure if the system was optimised for Agile it woukd be substantially better.
The big question is, is worth it? I have a system that works very well, and is pretty simple. I'd either have roll my own system or use Predbat or,Victron DESS.
I think even a basic system where I export more in the peak 16:00 to 19:00 hours would payoff, that would be very easy to setup, just involves adding another schedule and target SOC to my current schedules.