Opinion polls don’t tell us anything. They only suggest, tease, taunt and, for party leaders (especially those who claim they “never read them”) inspire panic.

As we get ever-nearer to the election, all this will only become worse. Every blip and dip will be written up either as so and so bouncing back, or someone else slumping into crisis.

This is fun to do and makes lots of headlines, but is all rather misleading.

Here’s what I believe is a more constructive way of interpreting polls. It’s a graph showing the change in the monthly poll average of the main four parties over the past half year:

There’s no secret to calculating these figures. It’s simply the average of every poll published by every reliable polling organisation in any given month.

You can see how Labour’s lead over the Tories has narrowed gently, but that both parties appear to have dipped at the same time as support for Ukip has risen slightly.

For all the eye-catching polls of recent weeks (Ukip up to 20%, Labour and the Conservatives neck-and-neck), the popularity of all the parties seems not to have wavered all that much.

But if I trace the trends a bit further back, a more intriguing picture emerges:

For much of the past year, Labour maintained an impressively consistent poll performance. The party’s average remained between 38% and 38.8% for nine months in a row: no mean feat when you think of all the knocks and wobbles it experienced (including Ed Miliband’s “summer of discontent“).

Only in the last two months have things changed. The party’s average has dropped almost two whole percentage points from February to April. Over the same period Ukip’s average rose from 12.4% to 14%.

Is there a connection? Well, Labour’s rating dropped by a similar margin at roughly the same point last year, albeit from a much higher level (the party hit 42% in February 2013, but was down to 39.6% by April). The key thing is the party never recovered. The support it lost in the polls during those months never returned. Whether the same thing happens again this year will be worth watching.

Where might this support have gone? Not necessarily to their immediate rivals, the Tories. Their average hasn’t plunged so sharply this year. They were on 32.7% in February, rose to 33.2% in March, before falling back to 32.4% in April.

The party will also take heart from the gentle recovery that is suggested by its long-term trend in the poll averages: up 3.5 percentage points from May 2013.

For the Lib Dems, the most you can say is that they appear to have bottomed out. They seem to have an irreducible core of support that, come snub, smear or scandal, never dips or rises above 9-10%.

And Ukip? If I extrapolated the graph even further back to the start of 2013, they would be sitting below the Lib Dems. In January of that year their average was 9.8%, compared to 10.3% for the Lib Dems.

They then jumped up to an all-time high of 15.9% in May 2013 off the back of the local elections, before falling to 11.1% in December, then rising again. Both peaks have coincided with a drop in trends of support that have been more marked and more permanent for Labour than the Conservatives.

Again, what happens next will be worth watching.

I accept that poll averages, just like individual polls, don’t tell us a jot. But they can suggest things that one-off snapshots can’t: trends, swings, slow but steady movements of opinion, and a broader, perhaps more plausible political narrative that day-to-day headlines miss.

And it’ll be trends, rather than individual polls, that I’ll be paying particular attention to on this site throughout the next 12 months.

I’ll focus on patterns and developments that emerge over time – such as how it’s now a month and a half since the Tories got anything higher than 35% in any poll (25 March, to be precise); and that Ukip hasn’t polled anything lower than 11% since – well, fancy that – almost exactly the same day: 26 March.

Those, as Rafa Benitez once said, are facts. Albeit facts based on things that don’t tell us anything. So remember: with polls it’s not so much only connect, but only infer.

Note: two indispensable sources of up-to-date news and data on individual polls are Anthony Wells, editor of the superb Polling Report website, and Mike Smithson of the excellent Political Betting.