More leaflets = more votes!
I first heard this irritating but accurate one-liner from Gavin Barwell, when he was Campaign Director at CCHQ. My colleague, Chris Scott, also recites the mantra, as do I. Now I have proof.
Over the last few days, along with my 'resident statistician' Dr John Hayward, we have been number crunching the Kent and SE Regional voting figures from 2 May - and particularly within the three constituencies where I am agent. We have charted and cross referenced votes by Division over the last four years, and cross tabulated them against individual ward / divisional and candidate's campaign actions.
Without divulging specifics I can now say that "More Leaflets = More Votes" is true (not that I personally ever doubted it).
Over the 14 County Council Divisions where I was agent, there is a direct and clear correlation between the number of 'points of contact' with an elector and the size of the swing. I have always been frustrated at candidates who claim "it makes no difference", or whenever I hear the tired words "if we do too much it will irritate people". Almost always those making such comments haven't knocked on a door or delivered a leaflet for a decade!
What is absolutely clear is despite the national tide in UKIPs favour and a number of issues testing the loyalty of the traditional wing of the Conservative Party; where a local candidate had a high profile doorstep based campaign, they were easily able to see off the threat.
By way of example; in our three top performing Divisions, the Conservative vote share (%) fell by less that 5% (and this is compared with 2009, which was a record year for us). In our worst three Divisions, the Conservative share fell by more than 15%. And can I share something else ? In the top tranche, our local teams had 3 x the number of contacts with voters as the bottom tranche.
So next time you hear a laissez faire candidate or branch Chairman claim "campaigning makes no difference" - tell them they are wrong!
You analysis sounds interesting but perhaps it is time for a proper split test on the matter. In one of your unwinnable wards perhaps you could split the electorate into three (randomised by Mosaic code, gender and if your council provides it, age).
ReplyDeleteYou could then deliver nothing to one group, strongly branded material to another and local newsy sort of stuff to another. Alternatively you could merely change quantity, none, some and loads.
These folks could then be measured by telephone survey or canvassing. Their marked register returns should also be checked to spot any correlation.
Thanks Anon, but we are ahead of you in two respects. Firstly, we don't have an unwinnable ward (which makes life tricky)! However, we do have a large two member division, where lack of resource meant we could only focus GOTV in five of the seven local authority wards which make up the Division. I deliberately left one ward with nothing but an election address, and in the other we delivered an EA and a GOTV leaflet. The other 5 had EA, GOTV, tellers and full polling day knock-up by door and phone. as soon as I can get the marked registers, I can see how the t/o of pledges differed from the five targeted wards compared with the one with partial GOTV and the one with no GOTV.
ReplyDeleteI have also sent 1,000 random surveys to recent C pledges in the Division with the best result and the Division with the worst result. The surveys are totally anonymous. I have asked the pledges to confirm if they voted Con on 2 May and if so, given various quantitative options to identify what, if anything, affected their vote and loyalty. I am keen to identify what specific factors (leaflets, doorstep, direct mail, personal contact, candidate's record etc)had impact on their VI.
Far too little work has been done locally to identify what actually works, so we (all parties) continue doing what we have always done, whilst adding new layers of campaign technology. In my first election (30 off years ago) we had an EA and a Last Minute leaflet. Now we have road specific in touches, EA, blue ink letters, personalised GOTV card, variable paragraphs. We do need to review what does and doesn't work.
Brilliant. Way ahead of me.
ReplyDeleteAfter finding out what works we need a way to get it out to our people (and not theirs). I know CCHQ have a Best Practise who should probably be given any results of testing but CCHQ's interaction with the associations and not great.
Some sort of central resource which is passworded.