The first lesson is that similar polls taken at similar times can produce quite different outcomes.
Such an occasion happened last week when the TVNZ Colmar Brunton poll had pretty much a stay-put outcome with National on 45 per cent, Labour on 42 per cent and both the Greens and New Zealand First above the 5 per cent threshold for list seats in Parliament.
Though National's not-so-new Leader, Simon Bridges, dropped a couple of points in the preferred Prime Minister stakes to 10 per cent at a time of maximum exposure via his party's conference, he and Sir John Key were able to point to National's steady party vote score as some sort of endorsement of his leadership.
This result meant Winston could quip that three-quarters of his own voters didn't like Bridges, but in general the commentariat swallowed his line and predicted no threat to his leadership.
How much different the comments and indeed the outcome might have been if an internal party poll taken at roughly the same time had taken the place of the published TVNZ Poll.
This particular poll had National on 39 per cent and Labour on 45 per cent, with the Greens and NZ First, like the TVNZ Poll, above the 5 per cent threshold.
READ MORE: Mike Williams: No Spygate but mole reveals new Muldoon wears dress
READ MORE: Mike Williams: Target on Gayford's forehead because he's an asset
READ MORE: Mike Williams: Jacinda effect opens wallets
This would have been interpreted as a near disaster for Bridges, with National shedding more than a tenth of its vote at a time when Labour's biggest asset was out of the ring and Simon Bridges was receiving maximum exposure.
This is not an unusual occurrence. I always judged that the TVNZ poll had a slight, perhaps 2 per cent, bias towards National and tiny differences in methodology can produce different results.
The lesson here is that polls will vary greatly and won't be great predictors of political outcomes in years ahead.
This lesson was best demonstrated by Massey University Academic Professor Claire Robinson, who in these very pages cited polls to confidently predict there would be a National Government after last year's general election. Whoops!
Under MMP, the fate of the smaller parties is of considerable importance in deciding election results and it is in this mathematical region that polls are least accurate.
The other poll that dominated the airwaves was the ANZ Business Outlook survey which showed what was generally interpreted as a dangerously negative mood in the business community and, like the year 2000, another "winter of discontent".
The second lesson is to always look for independent statistics to verify this kind of opinion-based finding.
Despite the negative outlook of the 350-odd self-selecting business people in the ANZ survey, the TVNZ poll showed little movement in the economic outlook indicator, with optimism down one point to 39 per cent while pessimism was steady on 35 per cent.
This seems to show that the dire outlook of ANZ's business panel isn't shared by voters in general.
It's a fair assumption that the bulk of the ANZ respondents would be National voters as any analysis will tell us that people on CEO-level salaries generally support the right-wing side of politics.
It's also fair to believe that their responses are coloured by their political preferences because in the real world these captains of industry are behaving with great confidence.
A solid measure of real business confidence is the fleet vehicle market and this tells a very different story.
The bulk of new cars in this country are sold to business and any real weakening of business confidence would show up in vehicle sales.
With new-vehicle sales heading a record year, the ANZ survey respondents' actions contradict their survey responses.
The Labour-led Government should treat these findings with a large grain of salt and stay the course.
• Mike Williams grew up in Hawke's Bay. He is CEO of the NZ Howard League and a former Labour Party president. All opinions are his and not those of Hawke's Bay Today.