The perfect Conservative Association would canvas every voter in its constituency between general and local elections.  Or would it?

The proposition sounds like a no-brainer, but opinion within the Party is divided.  Some say that all knocking on every door achieves is to find opposition voters as well as Tory ones – thus motivating the former as well as the latter to turn out in due course.

They prefer a mix of surveys, the use of such consumer classification systems as Mosaic, and targeted canvassing to identify committed and potential Conservative voters.

Others insist that there’s no substitute for door-knocking or telephone canvassing in an attempt to find every voter.  This morning, we want to give this disagreement some context – in the form of the three recent by-elections: one won, two lost.

It’s worth pointing out at the start that we have been here before.  In the wake of the 2017 general election, Mark Wallace wrote a magisterial audit of the campaign for this site called the Rusty Machine.

The second part of it featured an analysis of how activists were sometimes sent to knock up opposition voters rather than Tory ones – because they were armed with the wrong information.

It’s been reported that this happened all over again in the Batley & Spen by-election – and that, since Labour held the seat by only 300 votes or so, the error was decisive.

ConservativeHome has heard similar claims about the contest in Chesham & Amersham.  Having talked to some of those involved in all three by-elections, we believe that the following five lessons apply for future ones.

1. Make a retail offer

It was put to us that the core activist team in Hartlepool was very small – “no more than 20 people”, we were told (though it was supplemented by visits from Tory MPs, the Northern Research Group and others.

And that the pledge base was “no more than a few thousand at the most”.  Neither of these shortcomings turned out to matter much because the Party had a retail offer: Ben Houchen.

What seems to have made a big difference in Hartlepool was Houchen’s reputation for delivering for voters in next-door Teesside – the factor that drove his stonking win in its recent Mayoral election.

“People thought: we’ll have a bit of that,” one source said – painting a knowing, transactional view of Hartlepool’s voters.  “So they were prepared to give us a try in exceptional numbers”.

In Chesham & Amersham and Batley & Spen, by contrast, there was no similar offer – but, rather, campaigns based on delivering national pledges in a local context; and of representing the candidates as campaigners for local people.  The same traction wasn’t achieved.

2. Know where your voters are

If finding, motivating and getting out supporters was important even in Hartlepool, where the political winds were favourable, it was even more so in the other two contests.

“There was nothing at all on Vote Source for Batley & Spen,” one source said, claiming that some of the voter information available was out of date – stretching all the way back to the 2001 general election.

On election day in Chesham & Amersham, by contrast, there were 17,000 people to knock up.  But even so, ConHome is told that a significant part of that total was based on old constituency data which preceded the general election of 2017.

One senior figure pointed out that, in “safe seats”, as Chesham & Amersham was previously considered to be, motivation to find Tory voters at general elections is limited, since they will turn out anyway in numbers sufficient to win.

And that in constituencies like Batley & Spen, last represented by the Conservatives in 1992, activists face general elections knowing that their seats have not been designated as targets – so their incentive to campaign will also be limited.

In Hartlepool, a combination of Houchen’s reputation and its Brexity history meant that a low pledge base turned out not to matter.  But it usually does.

3. Don’t rely on predictive systems

One source from Batley & Spen insisted that reports of activists being sent to knock up the wrong voters, on the basis of wrong returns from surveys or mistaken voter identification from Mosaic, are wide of the mark.

“The fact is that voters change their minds, and sometimes forget they’ve done so,” he said.  “The Conservative pledge near the start of the campaign isn’t necessarily one at the end”.

“This was a complex, fast-moving election, and the claim that it was lost because of misentered data is for the birds”.  We were also told that Mosaic performed well in Chesham & Amersham.

However, others insist that this account is simply wrong – that not all the voters who claimed when knocked up in both elections never to have been considering voting Tory at all will have been forgetful, untruthful or self-deceieved.

One canvasser told ConHome that the predictive system “was proved wrong on almost every door I knocked on – an NHS administrator here, a college lecturer there: a series of high income socialists”.

Common sense suggests that in an era when voters are restless, predictive tools will sometimes be wide of the mark, can’t take late swing into account and may be of limited use where factors other than income and wealth are concerned, such as ethnicity.

4. There’s no substitute for good campaigners

“For a start, the team were less concerned with obtaining thousands of pledges and more interested in meeting and speaking to people,” Adrian Lee wrote recently on this site, in the wake of a canvassing session with Andrew Rossindell’s team in Romford.

Rossindell is a legendary campaigner, but does what works for him in between elections – that’s to say, conversations with voters – work as well as well by-election, especially if they’re held on polling day itself?

Mark’s 2017 account explained the problems inherent in attempting to pick out “a few people in the same street…which is very time-consuming even if the targeting is correct”.  He also reported that the canvassing scripts for the purpose were “less than ideal”.

One source from the Batley & Spen campaign said that talking to possible Tory voters with a script election day had its part to play, and suggested that those undertaking this work didn’t always do it effectively.

At any rate, there’s clearly no substitute for experienced campaigners and canvassers on the doorstep.  On which point, two Conservative MPs complained to ConHome that the Parliamentary Party wasn’t properly deployed.

“There’s not much point in travelling all the way from Westminster to Hartlepool only to deliver leaflets,” he said.  It should be added that voters can’t always or even usually be found by knocking on door, especially during working hours.

5. Pool resources

There are reasonable objections to much of the above.  Even assuming it is desirable to canvas every voter between elections, where are most Associations to get the activists from to do so?

As for knowing where your voters are, won’t most local Associations’ ability to capitalise on information be limited – again, because of the limited number of local members and helpers?

Since this is so, what practical alternative is there but to target many potential supporters with literature, and to use surveys and Mosaic-type systems in so doing?

These questions have a point, but there’s a missing answer – or part of one.  Namely, activists from non-target seats helping those in target seats between elections.

It happens, of course.  But it’s a pity that the problems that toppled RoadTrip (later RoadTrip 15) have frightened the Party off from organising programmes of structured constituency help on a similar scale.

Admittedly, such ventures might be no use at come a by-election, if the seat in question hadn’t been targeted for help.  But moving activists into target seats, within limits, must be the right thing to do in elections and between them.

Finally, some points to wrap up with.  Even the best-oiled campaigning machine won’t deliver in by-elections if voters are determined to protest against the Government.

The Matt Hancock affair is likely to have made at least as much difference in Batley & Spen as mistaken information.  There is a wider conversation to be had about the relationship between campaigners and local elections. And about the role of social media.

Above all: there’s no substitute for activists – be they members, deliverers or helpers.