Campaign 08 Blog logo Campaign 08 BlogPolitics meets the social web
Political Ads

What Election Games Get Wrong About Buying Political Ads

In most election simulators, buying an ad is a shopping trip. You open a panel, drag a slider from zero to some number of dollars, pick a state, choose attack or positive, and press confirm.

A laptop showing an election strategy game map beside a printed station media buy order and a notebook of market names

In most election simulators, buying an ad is a shopping trip. You open a panel, drag a slider from zero to some number of dollars, pick a state, choose attack or positive, and press confirm. Next turn your support in that state moves by a point or two. I have put several hundred hours into games built on this loop, from The Political Machine through Democracy 4 and President Infinity, and the loop is genuinely fun.

It is also the part these games get least right, and not in the way the usual criticism assumes. People complain that simulators make money too powerful. Having spent a lot of time reading public files and ad libraries alongside playing them, I think the sharper error is different: the games model money as the constraint when the real constraint is usually time, inventory and the thirty people who have to approve the script.

The slider that does not exist

No campaign buys advertising at a posted price. Digital inventory on the large platforms sells through an auction, so what you pay for a thousand impressions depends on who else wants to reach the same people on the same evening. In a quiet August week a campaign might reach a persuadable voter in a mid-sized market for 8 to 12 dollars per thousand impressions. In the last ten days before a close election the same audience can cost three or four times that, because every committee, party and outside group is bidding for the identical narrow slice of people.

Broadcast television works on a different and stranger rule. In the United States, a candidate's own committee gets the lowest unit charge, meaning the station must sell the candidate time at the cheapest rate it gave any advertiser for that slot in the preceding period. Outside groups get no such protection and routinely pay several times more for the same thirty seconds. That single asymmetry shapes who advertises where, and no simulator I have played includes it.

A game slider implies you can always convert dollars into reach. The real answer in late October is often no. A station has a finite number of spots in the 18:00 news, and once they are sold, your money has nowhere to go. Campaigns reserve inventory months ahead for exactly this reason, which is a decision with no representation in a turn based game.

What the games compress, side by side

Here is the comparison I keep coming back to when a simulator feels too clean.

In the gameIn practiceWhy the difference matters
Ad takes effect next turnScript, legal review, production and trafficking take 4 to 10 daysResponse ads to a late attack often miss the window entirely
One price per stateAuction pricing by market, daypart and audience segmentTwo campaigns can pay very different amounts for the same viewer
Attack or positive togglePersuasion, mobilisation, prospecting and retargeting, usually running at onceMost spending aims at turnout, not at changing minds
Unlimited inventoryFinite spots, reserved months in advanceLate money gets pushed to worse slots or to digital by default
No paperworkDisclaimers, FCC public file entries, platform verificationCompliance decides what can run and how fast
Spending is privateBroadcast buys are public, often within 24 hoursOpponents read your buys and know where you think you are weak

That last row is the one I wish a game would build around. A campaign's media buy is a public statement of its own internal polling. When a committee suddenly reserves time in a market nobody expected, reporters notice within a day, and the story becomes the buy itself. Any disclaimer on a spot tells you who paid, which is where to start if you want to trace one yourself. Our walkthrough on reading a political ad disclaimer covers the mechanics.

Why the simulators still teach something real

I am not arguing these games are useless. Two of their mechanics are closer to the truth than most commentary about advertising.

The first is diminishing returns. Good simulators make the tenth million in a state worth far less than the first, and that matches the research reasonably well. Studies of television advertising in presidential races have repeatedly found effects that are measurable, short lived and small, often fading within a week or two. A game that lets you buy a state outright would be worse, not better.

The second is opportunity cost. Every dollar you put into advertising in Democracy 4 is a dollar not spent on field organisation or policy, and that trade-off is the actual job. Campaign managers describe their work as a sequence of allocation decisions under a deadline, which is exactly the shape of a strategy game turn.

Where I would push a designer is on the calendar. Make the player commit broadcast reservations in July with incomplete information, then let the race move. Charge a penalty for late buys. Add a four day production lag. Those three changes would make a political simulator feel dramatically more like the public record without adding a single new menu.

Reading a real buy the way you read a game screen

You can check all of this yourself, and it takes less time than a single game session. Station buys for broadcast are filed in the public inspection file maintained by the regulator, with the rate, the number of spots and the dayparts. Platform ad libraries show creative, rough spend bands and who was targeted, though the detail varies by platform and by country.

What to look for, in order. The market, because that tells you which voters the campaign has decided matter. The daypart, because daytime cable buys cheaply and skews older while streaming skews younger and costs more per impression. The flight dates, because a buy that starts the week before early voting opens in that state is a turnout play, not a persuasion play. Then the creative itself, where the first five seconds carry almost all of the message that most viewers will ever receive.

Totals give you a sense of scale that no game conveys. Tracking firms put the 2024 United States cycle somewhere above ten billion dollars across all offices and media, with a single competitive Senate race occasionally passing a hundred million. For a feel of how the published numbers are organised, our guide to the ad libraries that track campaign spending lays out what each source actually covers.

The mechanic nobody has modelled

Here is my favourite gap, and it is the reason I keep returning to this subject. In every simulator, an ad reaches a population. In reality, an ad reaches a list.

Modern buying starts with voter files, matched to devices and to audience segments, so a campaign is not addressing Ohio. It is addressing 140,000 specific households it has scored for persuadability, and it can stop addressing any of them the moment a model updates. A game that represented this would need the player to maintain data quality, pay for matching, lose coverage when a platform tightens its rules, and accept that one in five of the people they bought have moved house. That is a less satisfying slider and a far more accurate one.

It also explains why advertising inside games themselves has become a line item rather than a novelty, which is a subject in its own right in our piece on political ads in video games.

Frequently asked questions

Which election game handles advertising best?

Democracy 4 models trade-offs and public opinion drift most convincingly, while President Infinity is better on the geography of a race. Neither models auction pricing or inventory scarcity, so both leave out the two factors that dominate a real October.

Do political ads actually change votes?

Measured effects on persuasion are small and fade quickly, which is why so much spending now targets turnout among people who already agree. That does not make advertising pointless. In a race decided by a point, a small and temporary effect is enough to matter.

Can I see what a campaign paid for a television spot?

For broadcast stations in the United States, yes. The public inspection file lists the order, the rate and the spots, and it is online and free. Cable and streaming disclosure is much thinner, which is one reason spending has moved in that direction.

Why do the same ads follow me after I have already voted?

Audience lists go stale and almost nothing tells an advertiser you have voted. Early voters commonly keep receiving mobilisation messaging for days. More on how the targeting is assembled sits across our political ads coverage.

MA
Martin Ashby

Martin has played political strategy and simulation games for twenty years, from Diplomacy by mail to modern election simulators. He writes about what those games get right about campaigns and what they flatten.

More posts by Martin

More in Political Ads

01doorganico 084goad 3hot8 4465456 after5internprogram ahandfulofblessings ahmadpc ahomcybersecurity allodialresources annodex antiquesandperiodcollectibles aotc asianworldkashmir beedlectrics bloggertemplatefree bluefeatheragency bodyweightsolution carfixadvisor catstreetcollective cbstvsports cnfanspurchasing committeeof500yearsofdignityandresistance covertopsdjs cs-kaiyuangame cuestacomunicaciontotal cxoutlet dancebetgame deskgearhub destinyvillastay detectmagazine di33 diamondsteelbuildings donmo dyanadropship emarsysfinance emimattl escapechan eskitarih evautoplay everythingweddings-orlando farmzone feedviewerapp foxstudiosscratch gamevui123 gleasongrade greetingsday hacklabasu haskell iw-kaiyun jagoanproperti jilidskripsi joltbusinessfinance juegodepokergratis juegospucca kaeru-festival latestmehndidesigns lestanzediottaviano litemagazines localcommunityhero lustrsazan mariashebets menswheels newwebstechs1 nikkit ootheday oxigenvpn plurfitness prefetial probudgetcarpetcleaning psychicmrsmartin pyaten qvantor rejoignezlogis rexzbot riselayouts ruchin rwxqsgd5 sakaltime shibiaowang shop-bot skrtips spotlesshaven strcoding tetapatrium topselller uneplanteartificielle vipertechcr webtech1223411 wo-kygaming yamaha-mekar yeknava yoybuysheets zenunwa yamaha-mekar jilidskripsi ankertoto ankertoto ankertoto ankertoto