Freshwater Strategy And The Liberal Party: A Post-Mortem On The 2025 Polling Failure

Thursday, 6 November 2025, 8:12 am
Article: Hugh Grant

The 2025 Australian federal election saw one of the wildest polling failures in recent memory. The Liberal Party's internal pollster, Freshwater Strategy, completely misread the mood of voters, even after pocketing $1.5 million for their work.

Freshwater predicted a tight race, hinting at possible Coalition gains and suggesting Labor would scrape by with around 76 seats. The reality? Labor stormed home with 89 seats, while the Coalition limped to just 40.

This polling meltdown cost Peter Dutton his own seat of Dickson after 24 years in parliament. It also exposed some deep flaws in the way modern campaigns lean so hard on data.

Freshwater's final polls had Labor ahead by just 1.5 percentage points. In the end, the two-party preferred result hit 54.8 percent for Labor and 45.2 percent for the Coalition—a miss of almost four points, costing the Coalition dozens of seats.

Liberal Party insiders have started openly questioning whether dodgy polling led to bad campaign calls. Industry folks are also wondering how such basic mistakes happened, given all the talk about “sophisticated” research and the money spent.

Background of the Freshwater Strategy Polling Controversy

Freshwater Strategy burst onto the scene in 2022, quickly landing contracts with big media and major political parties. Their quick rise—and then their crash during the 2025 federal election—kicked off a lot of finger-pointing inside the Liberal Party.

Origins of Freshwater Strategy

Freshwater Strategy launched in 2022, hiring staff from the Liberals’ old polling firm. They pitched themselves as the next generation, bringing in seasoned political researchers and strategists.

Dr Michael Turner runs the show as chief strategist and pollster. He’s got a hefty resume, with experience in UK, Australian, and New Zealand campaigns.

The timing of Freshwater’s launch lined up with a shake-up in Australia’s polling world. Several old-school pollsters had come under fire after blowing previous election calls.

A bunch of ex-Crosby-Textor folks joined the founding team, giving Freshwater instant credibility in conservative circles. That helped them land early contracts.

Previous Political Polling Involvement

Freshwater built its name on a string of accurate state election calls before 2025. They got Victoria, New South Wales, Queensland, and Western Australia right, or close enough.

Their biggest win came during the Voice to Parliament referendum. Freshwater nailed the result, showing they could handle tricky political analysis.

Dr Turner’s international work across several countries helped set Freshwater apart from local-only pollsters.

Their method leaned heavily on focus group research, not just the usual number crunching. They ran sessions with about 10 people in key seats, several times a week during the campaign.

Partnerships with Media Outlets

The Australian Financial Review became Freshwater’s main media partner for public polling. Their regular two-party preferred numbers ran in the paper’s election coverage.

Just before the 2025 election, Freshwater’s AFR poll said Labor would likely form a minority government. Their numbers put Labor at 51.5 percent, Coalition at 48.5.

But the real result was a Labor landslide. Final counts put Labor at 54.8 percent, Coalition at 45.2.

Freshwater didn’t just do polling for the media. They also offered detailed analysis and commentary on political trends for digital newspapers and political newsletters.

Dr Turner frequently wrote opinion pieces for the AFR, explaining how they did their polling and why they made the calls they did. It helped raise the company’s profile in political media circles.

The Liberal Party's Reliance on Freshwater Strategy

The Liberals went all-in on Freshwater for the 2025 campaign, dropping about $1.5 million on polling and focus groups. This hefty spend shaped the Coalition’s strategy and campaign direction.

Financial Investments in Polling

The Liberals handed over $1.5 million to Freshwater for polling. That’s one of their biggest spends on polling in recent campaigns.

Freshwater ran tracking polls every other night in 15 key marginal seats. Each poll surveyed 1,200 voters to get a read on the battlegrounds.

Focus groups ate up a big chunk of the budget. These sessions, about 10 people each, ran multiple times a week in target seats. The Liberals leaned on focus groups more than ever before.

Dr Mike Turner oversaw the polling. His team, formed in 2022, included veterans from Crosby-Textor.

Party insiders argued the high cost was worth it for the depth of the research. After the election, though, plenty of donors were fuming about the value for money.

Role of Andrew Hirst and Campaign Infrastructure

Liberal Party director Andrew Hirst managed the Freshwater contract. He controlled how the polling data was shared and interpreted during the campaign.

Hirst kept polling info on a tight leash. Only Peter Dutton, his office, and party directors saw the full data. Many MPs who complained about the polling never actually saw it.

Using Freshwater’s data, Hirst figured they’d end up with seat numbers in the low 60s. He expected Labor to scrape a slim majority, not the blowout that happened.

The campaign’s infrastructure revolved around this polling. They decided where to send resources based on Freshwater’s calls about which seats to chase or defend.

Hirst isn’t talking about the contract details. The Liberals plan to ditch Freshwater when their deal ends in June 2025.

Historical Context of Liberal Campaign Tactics

The Coalition’s approach marked a shift from old-school campaign tactics. In past elections, Liberals balanced polling with broader strategy and on-the-ground smarts.

Freshwater replaced Crosby-Textor as the main pollster. That was a big break from the relationships that guided many previous campaigns.

They leaned much harder on focus groups than before. Traditionally, the Liberals favored big, broad polling over small group chats.

Peter Dutton’s public confidence—talking up “big surprises” and comparing 2025 to 2019—came straight from Freshwater’s numbers.

Strategic decisions, from spending to messaging, were driven by the polling. Relying so much on one research source was a big shift for the party.

How Polling Data Shaped 2025 Election Campaign Decisions

The Coalition’s dependence on Freshwater’s polling set off a chain of strategic choices that shaped their 2025 campaign. Their interpretation of the data directly affected where money and effort went, and who they tried to reach.

Marginal Seat Analysis Methods

Freshwater used seat-by-seat modeling that mixed past voting patterns with demographic shifts. They sampled 800-1,200 voters per marginal seat.

They tracked three main things. Primary vote tracking measured support for the big parties each week. Preference flows estimated two-party preferred outcomes. Issue salience scoring ranked what mattered most to voters in each seat.

The Coalition got detailed breakdowns for 35 marginals. Each report included voter movement and demographic splits. Freshwater’s models predicted swings based on how much local campaign activity there was.

Strategists used these reports to pick “winnable” seats and decide which ones to defend. The data suggested some outer suburban demographics were drifting toward the Coalition.

Influence on Coalition Resource Allocation

Polling results drove big funding calls. TV ad budgets went to markets where Freshwater saw close races. Radio spending followed the same logic.

Staff deployment matched polling confidence. Safe seats got less attention, while marginals with positive trends got more campaigners and candidate time.

The party spent $2.8 million on targeted digital ads, using Freshwater’s demographic analysis to guide them. Social media campaigns tailored messages to age groups and regions.

Ground campaign resources shifted three times during the campaign. Each time, new polling spurred a reallocation to seats where things looked promising. Some traditional strongholds lost support as resources moved to new targets.

Interpretation of Voter Demographics

Freshwater’s demographic analysis shaped the Coalition’s messaging. They identified suburban families and regional workers as key swing voters.

Polling said economic worries topped the list for these groups. Healthcare and education mattered less in marginals, while climate change showed up differently by age and region.

The Coalition crafted separate messages for each group. Older voters heard about economic stability and security. Younger people got tech and opportunity themes.

Campaign tactics shifted as feedback came in. Town halls targeted regional areas where polling was strong; digital-first strategies aimed at urban voters moving toward the party.

Gender-based messaging popped up after Freshwater flagged different issue priorities between men and women in key seats.

Key Players: Peter Dutton, Scott Morrison, and Leadership Decisions

Peter Dutton and Scott Morrison made big calls that shaped the Liberals’ approach to the Voice referendum. Polling pointed one way, but leadership went another.

Peter Dutton's Response to Internal Polling

Peter Dutton took over as opposition leader while internal polling showed complicated attitudes toward the Indigenous Voice proposal. Coalition pollsters found that supporting the Voice might actually help the Liberals.

Even with that advice, Dutton decided to oppose the Voice referendum. He went against recommendations from long-time Coalition pollsters.

Dutton’s team focused on raising doubts about the Voice. They argued it would divide, not unite. He stuck to this message, even as polling suggested a different approach could work better.

Dutton’s main argument centered on constitutional concerns. He claimed the Voice would give Indigenous Australians special rights. That became the core of the No campaign.

Dutton’s decision-making seemed to put ideology ahead of polling. Former pollsters confirmed they’d advised backing the Voice, based on their data.

Scott Morrison's Previous Campaign Approach

Scott Morrison’s time as leader set patterns that influenced the Voice referendum strategy. As PM, Morrison saw similar polling about Indigenous issues and voter preferences.

Morrison tended to play it safe on Indigenous policy. He supported some recognition but dodged big constitutional changes. That shaped expectations inside the party.

Morrison’s decision-making leaned on political calculations. He often took positions he thought would please the Liberal base, not necessarily the broader public.

Even after stepping down, Morrison’s approach to the Voice issue stuck with the party. It influenced how they planned their referendum strategy.

Just like Dutton, Morrison ignored polling that said backing the Voice might actually play well for the Liberals. This habit of brushing off professional advice became a real feature of party leadership on the issue.

Polling Methodologies and Analytical Failings

Freshwater’s polling disaster came down to basic flaws in their methods and analysis. Their so-called “evidence-led” approach just didn’t capture what voters were feeling, and mistakes in weighting and preference flows made things worse.

Evidence-Led Versus Traditional Methods

Freshwater pitched themselves as using an “evidence-led” approach, different from old-school polling. Dr Michael Turner’s team claimed they could turn raw data into strategic gold.

They focused on marginal seats instead of big national samples. Tracking surveys ran every other night in 15 key marginals.

What set them apart:

• Marginal seat focus, not national
• Emphasis on “strategic insights” over just the numbers
• Mixed focus group data with stats
• Regular tracking, not just snapshots

This method just didn’t pick up the late swing to Labor. Broader, more traditional polling probably would’ve caught it.

Weighting Mistakes Linked to the Voice Referendum

Freshwater made some big errors in how they weighted their data. They didn’t properly account for how recent political events, like the Voice referendum, shifted voter behavior.

These weighting mistakes threw off both primary vote estimates and two-party preferred numbers. Their final poll showed Labor at 51.5%, but the real number was 54.8%.

That’s a four-point miss for Labor. In seat terms, that’s a disaster—Freshwater predicted 76 Labor seats, but Labor ended up with 89.

Their models just couldn’t factor in how the referendum affected voting patterns.

Survey Technology and Data Analysis Techniques

Freshwater ran regular surveys across target seats, costing the Liberals $1.5 million during the campaign.

Their tracking surveys polled 1,200 voters each cycle. But that wasn’t enough to catch subtle but important shifts in the final weeks.

Technical problems included:

• Sample sizes too small for marginals
• Bad timing on survey cycles
• Weak demographic controls
• Poor late-movement detection

Their tech didn’t pick up how well Labor’s “protecting essential services” message was landing. That oversight played a big part in their polling failure.

Estimation Errors and Preference Distributions

Dr Michael Turner called out preference distribution calculations as one of the three big failure points in Freshwater's methodology. The company seriously underestimated how minor party preferences would flow to Labor.

Freshwater's models leaned too heavily on Coalition-friendly preference flows. This mistake just made their primary vote errors worse, leading to wildly off-base seat predictions.

They predicted Coalition seats in the low 60s, sometimes even hinting at the high 60s. In reality, the Coalition only managed 40 seats—a 20-seat underestimate at best.

Preference distribution mistakes really stung in close contests where minor party votes tipped the scale. Freshwater missed how well Labor scooped up preferences from independents and smaller parties.

The Collapse: What the Polls Predicted vs. Actual 2025 Results

Freshwater Strategy's polling predictions fell apart when compared to the actual 2025 federal election results. Their forecasts suggested a tight race with possible Coalition gains, but Labor ended up crushing it—way beyond what anyone expected.

Forecasts Published in Media Outlets

Freshwater's final poll in The Australian Financial Review said Labor would scrape together a minority government with about 76 seats. That forecast painted the election as a nail-biter.

They projected a modest swing to Labor. Their two-party preferred estimate put Labor at 51.5 percent and the Coalition at 48.5 percent.

Liberal Party director Andrew Hirst expected the Coalition to land in the low 60s based on Freshwater's numbers. The research never showed the Coalition getting into the 70s.

Peter Dutton referenced these internal figures all campaign long. He kept suggesting the Coalition could pull off another 2019-style upset, like when Scott Morrison won.

Discrepancies in Two-Party Preferred Estimates

The real election results exposed huge errors in Freshwater's calculations. Labor scored 54.8 percent of the two-party preferred vote, leaving the Coalition with 45.2 percent.

That’s almost a four-point miss. Freshwater had Labor ahead by only 1.5 points, but the real margin blew out to 9.6.

Freshwater calculated a swing of just 0.6 percent since 2022. The actual swing was much bigger and wiped out Coalition hopes in seat after seat.

Dr Michael Turner from Freshwater Strategy later pointed out three main mistakes. Among them: overestimating Labor voters switching to the Coalition and messing up preference distribution calculations.

Unexpected Losses for the Coalition

Labor's final seat count hit 89, blowing past Freshwater's projection of 76. The Coalition ended up with just 40 seats, not the low 60s as forecast.

Peter Dutton lost Dickson after 24 years in parliament. The Liberals didn’t even bother polling Dickson in the final weeks—they figured it was safe.

The Coalition’s primary vote fell harder than any poll had suggested. Labor’s campaign about protecting essential services landed better with voters than Freshwater’s surveys picked up.

Liberal Senator Johnathon Duniam called the polling "way off the mark" and "totally out of line with all published polling." The Liberal Party decided to end Freshwater Strategy's contract.

Aftermath and Repercussions for Australian Political Polling

The Freshwater Strategy disaster set off immediate contract termination and a wave of industry criticism. Inside the Liberal Party, MPs and donors were furious, and the polling failure forced everyone to rethink how campaigns use internal research.

Termination of the Freshwater Strategy Contract

The Liberal Party wasted no time cutting ties with Freshwater Strategy after the polling debacle. The party prepared to terminate the firm's contract when it expires in June 2025, according to senior Liberal sources.

Andrew Hirst had banked on a seat count in the low 60s from Freshwater's data. The actual result—just 40 seats—was a brutal shock that cost the party millions in wasted resources.

Liberal MPs were livid about the $1.5 million spent on what turned out to be bad intel. One senior figure called the internal polling "all bullsh*t" once the results sank in.

The contract termination sends a message—maybe overdue—about accountability in political consulting. The party’s move highlights bigger worries about transparency and reliability in the polling world.

Michael Turner's firm ran tracking polls every second night across 15 marginal seats. Even with all that coverage, their methods just didn’t pick up the real mood of voters in those final, crucial weeks.

Industry and Internal Liberal Party Reactions

Furious Liberal Party members blamed Freshwater Strategy for misleading the campaign and giving candidates a false sense of hope. Senator Johnathon Duniam didn’t mince words—he said the polling was "way off the mark."

Donors threatened to "close their chequebooks" after the campaign flopped. The financial fallout went deeper than just this election, making future fundraising a lot tougher for the party.

A big post-mortem, likely led by someone like Brian Loughnane, was on the cards. The review would dig into what went wrong and how bad polling shaped strategic decisions.

The disaster sparked a lot of talk in the industry about polling standards. Some political watchers started wondering if parties had gotten way too reliant on shaky internal research.

Freshwater Strategy's reputation took a major hit in Australian political circles. Other polling firms suddenly found themselves under the microscope too, with everyone demanding better methods and more honest accuracy claims.

Impact on Future Campaign Polling Practices

The Freshwater disaster threw a harsh spotlight on some big weaknesses in modern polling. Younger demographics proved "notoriously difficult to poll"—let’s be honest, most of them just ignore calls from unknown numbers.

Political parties are starting to rethink how much they pour into internal polling. Losing $1.5 million really drives home how risky it is to lean on just one polling source.

Campaign strategists now feel the heat to mix up their intelligence sources. If you only trust one firm’s numbers, you’re asking for trouble.

The whole mess also revealed how tricky preference flow calculations and late-swing detection can get. These issues call for sharper sampling methods and smarter data crunching.

I’d bet future campaigns will push a lot harder for transparency in polling methods and how folks calculate margins of error. The Freshwater case makes it clear—bad assumptions can ripple through everything.

Political consulting firms can’t just coast anymore. Clients want real explanations for how they come up with their numbers.

© Scoop Media