I was quite shocked about what they said about e-voting security:

• The level of implementation and security risk involved was significant and unacceptable.
• Although the MoJ undertook its own quality assurance through a security audit, this took place far too late ... which meant that, realistically, there was not enough time to make any significant changes following the audit.

Supplier performance was poor:

• we have significant concerns about the overall robustness of the procurement framework established by the Department for Constitutional Affairs ... it is clear that some suppliers underperformed ... This raises serious questions about their qualifications for being on the framework. The local authorities believed that the procurement process had involved a higher level of testing and investigation than was actually the case.

Sounds like if it hadn't been small scale local elections something could easily have gone very wrong. But on the small scale it did work actually work, despite all too short timescales.

E-voting take up was rather low, in part due to pre-registration requirements, so costs per actual e-vote were shockingly enormous. The costs aren't conveniently summarised, but delving into each detailed report it seems that the cost per e-vote in the trial was in the £100 to £300 range. I've attempted to summarise costs per voter / registered e-voter / internet-voter / telephone-voter below, for each of the 5 e-voting trials:

1. Rushmoor: £27.26 / £78.43 / £137.00 / -
   
2. Sheffield: - / £50 / £110 / £425
3. Shrewsbury&Atcham: £63 / £625.10 / £348.71 / £348.71 (advance voters cost £847.18 each)
4. South Bucks: - / £113.14 / £239.24 / -
5. Swindon: £8.33 / -/ £102.50 / -

So at great expense we've learnt some fairly obvious lessons. I think politicians should stop pushing this now, except perhaps in one permanent trial area where there is a prospect of gradually refining the technology and voter acceptance.