How much does personalisation increase response?

I'm talking about increasing response through personalisation at an upcoming Redeye Email marketing workshop, so just thinking through the variables in response to a question from an IDM graduate who asked:

I sat in on an agency briefing this morning re: email marketing to a number of different segements (trialists/registrants/subscribers etc) to promote new section of a web site, and the agency guy said that in just the past few months they had noticed that using personalisation on emails (and acknowledgement in copy of which segment they came from) was no longer a good idea, and could depress response....???

This seems to go against everything else I have been hearing - is this what you're finding too..??

Dave's answer:

Personalisation is a balancing act between these variables which all email marketers control. If you look at campaign response this way, you can see that personalisation can quite easily decrease response if you get the campaign execution wrong:

1. Relevance of offers (SHOULD  increase response)

2 Context (is the offer currently relevant to the subscriber, if so SHOULD increase response - this is why event triggered emails work so well)

3. Number of offers (MAY decrease or MAY increase response since more offers can lose focus of message but provide wider choice)

4. Form of personalisation (recognition of individual SHOULD increase response but over-familiarity MAY decrease especially in subject line)

This report has data on personalised subject lines that suggests it varies by country:

5. Targeting (more granular segmentation with a relevant offer and message SHOULD increase response, but only if the data is correct and the right offer for the target is selected).