Let me start by apologizing to all the nice subscribers I’ve gained whose kindness I repaid by immediately ghosting them. November and December have become busier than I anticipated, which is one of those good problems I suppose. At some point I’d like to write about Zohran’s win in the mayoral, the attacks he faced, and what it teaches us about opposition research, but that will have to wait. Today I’d like to share some quick thoughts on the entry of Rep. Jasmine Crockett into the Texas Senate race.

I don’t have fully formed thoughts on Rep. Crockett and what her candidacy means for Democrats’ chances in the Senate. You can find endless takes on that subject and I wouldn’t have anything original to add. What I did find intriguing, however, is what Crockett reveals about the modern political fundraising environment and its players.

A Blue Chip Client for Blue Chip Strategies

Crockett is famous for her media virality and preternatural ability to cast herself as a leading antagonist to President Trump. She’s genuinely a new media powerhouse. I mean, just look at whatever this is. And Crockett AI videos are a whole genre, just search for ‘jasmine crockett candace owens debate’ to get a baffling taste of ai narrated videos depicting arguments that never happened.

That all has translated into healthy fundraising: she went from raising $970,008 in the 2021-2022 cycle, which was her only contested federal election, to $3,114,949 in 2023-2024, to over $6.5 million so far in 2025. And that’s without reporting the fourth quarter and what is sure to be a formidable haul from her Senate announcement. Exponential fundraising growth is a serious achievement for any member of the House, and it’s surely one reason she feels confident she has a chance to win statewide.

As is the case for any candidate that interests me, I decided to poke around in her FEC reports. You can find a lot out about a candidate from how they choose to spend their money. She’s far from the first to spend donor money on nice hotels, so I’ll leave that coverage to the people who care more about that sort of thing. What caught my eye was her largest vendor: Blue Chip Strategies LLC. Over $850,000 this year, reportedly for “Advertising,” “Digital Consulting,” and “Fundraising Consulting.” In total, Crockett has paid Blue Chip $1.16 million in twenty-four separate payments dating back to February 2024.

That’s a lot of money for any vendor, though presumably much of it is covering the actual digital media buys themselves—ads on Facebook and Instagram, barraging people’s emails and text messages, the usual ingredients for online fundraising success. Sure enough, when you check Meta’s ad library, Crockett began posting fundraising ads in earnest as of May 2024, shortly after she first reported paying Blue Chip. According to Meta, Crockett’s various committees have spent $1.7 million on their platforms between December 2018 and December 2025, which lines up relatively well with the Blue Chip Strategies figure. The first ad of this era highlighted her spat with Rep. Marjorie Taylor Greene. Keep that in mind for later.

None of this is particularly interesting until I decided to look up Blue Chip Strategies. Only I found they have no web presence. A digital fundraising firm with no active website or social media account. I did stumble across “www.bluechipstrat.com,” but it’s private. A trip to the wayback machine shows it hasn’t been active in years and what was on there was the usual pablum about message development, crisis communication, and “Emmy Award Winning Story Telling.” Also this line from the CEO of the Harris X Poll: ““I’ve never seen ads score so well. They test off the Richter scale.”"

Checking their address in the FEC filings, you find a mailbox at Atlanta area UPS store. Looking the Georgia Secretary of State, it appears they used a registered agent to avoid listing principals.

Now I’m intrigued.

Blue Chip Strategies’ Clients

Blue Chip is listed as a vendor for 13 federal campaign committees since 2019. The breakdown is as follows:

1. Marcus for Georgia

• Candidate: Marcus Flowers (Georgia)

• Total Amount: $5,094,288.39

• Number of Payments: 69

• Purpose: Digital Advertising, Media Production, Technology Fees, Text Messaging, Reimbursements

• Timeline: 2021-02-24 to 2022-11-09

2. Jasmine for Us

• Candidate: Jasmine Crockett (Texas)

• Total Amount: $1,162,055.05

• Number of Payments: 24

• Purpose: Advertising, Digital Consulting, Fundraising Consulting, Online Advertising

• Timeline: 2024-02-28 to 2025-06-27

3. Booker for Kentucky

• Candidate: Charles Booker (Kentucky)

• Total Amount: $269,070.00

• Number of Payments: 7

• Purpose: Digital Consulting, Mileage Reimbursement

• Timeline: 2022-09-22 to 2023-08-29

4. Matt Lieberman for Senate

• Candidate: Matt Lieberman (Georgia)

• Total Amount: $218,476.58

• Number of Payments: 23

• Purpose: Fundraising Consultant, Media Buy, Media Production, Reimbursements

• Timeline: 2019-09-04 to 2020-09-08

5. Anna Kaplan for New York

• Candidate: Anna Kaplan (New York)

• Total Amount: $167,494.35

• Number of Payments: 8

• Purpose: Digital Advertising, Media Production

• Timeline: 2023-06-23 to 2024-02-15

6. Roland Gutierrez for Texas

• Candidate: Roland Gutierrez (Texas)

• Total Amount: $89,371.48

• Number of Payments: 5

• Purpose: Fundraising Consulting

• Timeline: 2023-10-06 to 2024-02-09

7. Youth Save Democracy PAC

• Purpose: A PAC dedicated to mobilizing young voters to elect Democrats to the House.

• Total Amount: $59,903.05

• Number of Payments: 6

• Purpose: Digital Advertising

• Timeline: 2023-08-08 to 2023-12-18

8. Sarah Riggs Amico for Georgia, Inc.

• Candidate: Sarah Riggs Amico (Georgia)

• Total Amount: $54,000.00

• Number of Payments: 5

• Purpose: Fundraising Consulting Services

• Timeline: 2020-02-04 to 2020-06-02

9. Georgia Federal Elections Committee

• Purpose: The federal account for the Democratic Party of Georgia.

• Total Amount: $50,000.00

• Number of Payments: 1

• Purpose: Fundraising List Purchase

• Timeline: 2022-11-17

10. Dana Barrett for Congress, Inc.

• Candidate: Dana Barrett (Georgia)

• Total Amount: $48,850.00

• Number of Payments: 13

• Purpose: Fundraising Consultant(s), Video Production

• Timeline: 2019-10-30 to 2020-08-31

11. Sites for Congress

• Candidate: Jeffrey Sites (Ohio)

• Total Amount: $40,892.00

• Number of Payments: 3

• Purpose: Digital Advertising

• Timeline: 2023-11-08 to 2023-12-01

12. Kaine for Virginia

• Candidate: Tim Kaine (Virginia)

• Total Amount: $33,195.30

• Number of Payments: 2

• Purpose: List Acquisition

• Timeline: 2023-07-18 to 2023-08-18

13. Web3 Forward

• Purpose: A PAC supporting candidates aligned with the development of Web3 and blockchain technologies.

• Total Amount: $15,447.68

• Number of Payments: 2

• Purpose: Editing, Media Production - Not For Independent Expenditure

• Timeline: 2022-06-14 to 2022-08-24

Overall Total: $7,303,043.88

The Marcus Flowers Connection

That Marcus Flowers number is bonkers. Flowers reported raising $16.9 million total for his now-forgotten, doomed campaign against Marjorie Taylor Greene. That’s the 6th most of any House candidate in the 2021-2022 cycle, behind only then-Speaker Kevin McCarthy, Katie Porter, Minority Leader Pelosi, Adam Schiff, and Majority Leader Steve Scalise. And nearly a third of it went to Blue Chip. Suddenly it seems like less of a surprise that hiring Blue Chip was timed with Crockett seeing the profit in beefing with MTG.

The above numbers shouldn’t be considered comprehensive: presumably there are state clients, 501(c)(4) clients, and others who wouldn’t report to the FEC.

But who is Blue Chip Strategies exactly? Luckily I’m not the first person to look into this. In a well-reported piece on how Greene and Flowers were both raking in donations online, the Atlanta Journal-Constitution noted “Bobby Kaple, a former TV news anchor and unsuccessful congressional candidate, and Michael Carcaise, a Colorado-based political consultant, started Blue Chip Strategies in 2019.”

Kaple was a former journalist at CBS46 in Atlanta who jumped into the primary against Republican Rep. Karen Handel in 2017. He came in third with 26.2% behind Lucy McBath (36.3%) and Kevin Abel (30.5%), who both advanced to a runoff. McBath would go on to win the primary and eventually the general.

Although ending in defeat, that primary appears to have served as something of an origin story for Blue Chip Strategies. A quick perusal of Kaple’s FEC filings shows that the campaign paid Michael Carcaise $34,956.23 between November 2017 and June 2018 for payroll, “Consultant - General,” and “Consultant - Campaign.”

While Carcaise keeps somewhat of a lower profile, Kaple has kept to his TV roots to become a talking head on local Atlanta TV. You can see his picture above.

Crypto Consultants

The FEC allows us to make a reasonable guess about how Kaple and Carcaise first came across Crockett. You’ll notice that Web3 Forward was listed as one of the clients for Blue Chip Strategies. Web3 was one of a collection of crypto PACs that popped up in 2022, spending $1.32 million in support of Jasmine Crockett in her campaign to replace Eddie Bernice Johnson in TX-30.

Here’s their statement shared with the Texas Tribune:

“Web3 Forward is supporting Crockett because of her commitment to give Texans the opportunity to build and use next-generation Web3 technologies that put power back in the hands of users, democratize financial access, and empower individuals to retake ownership of the internet and their own data.”

Notably, Carcaise was listed as the treasurer for Web3 Forward.

Web3 was funded by some of the biggest names of the era in crypto, per CNBC:

“Web3 Forward and Crypto Innovation are funded primarily by the GMI PAC, another super PAC that has seen massive donations from cryptocurrency leaders. The contributions include $2 million in January from FTX CEO Samuel Bankman-Fried. The group also got $1 million each in May from Marc Andreessen and Ben Horowitz, two managing partners at venture capital firm Andreessen Horowitz, records show. The firm’s portfolio includes cryptocurrency companies. Former Trump White House communications director and SkyBridge Capital founder Anthony Scaramucci donated $100,000 to Crypto Innovation in March, according to an FEC filing. GMI PAC’s founding board of directors includes CMS Holdings co-founder Dan Matuszewski, Framework Ventures co-founder Vance Spencer, and FTX Digital Markets CEO Ryan Salame, according to crypto news outlet The Block.”

In October 2022, Kaple pops up as a “Senior Advisor to GMI PAC” in two stories pushing polling about the impact of so-called crypto voters. GMI, a reference to the crypto aphorism “Gonna Make It,” also had a 501(c)(4) arm called GMI Policy Institute. And wouldn’t you know it, Bobby Kaple and Michael Carcaise are two of the three directors for GMI Policy Institute. The “Director and President” was Colorado resident Chase Goodwin, who during this same period shows up in the Atlanta Journal-Constitution story as Marcus Flowers’ campaign manager. Goodwin claims to have worked for John Barrow in the past, but I’ve already gone down enough of a rabbit hole.

GMI PAC has since dissolved, only to be replaced by a collection of super PACs called Fairshake, Protect Progress, and Defend American Jobs. According to Politico: “The crypto super PACs are being spearheaded by Michael Carcaise, according to a person with knowledge of the groups’ operations. Carcaise previously served as treasurer and as a strategist for the now-defunct GMI PAC.”

Combined, these three groups spent roughly $290 million in the 2023-2024 cycle supporting and opposing dozens of candidates. Bernie Moreno alone earned $40 million in support from Defend American Jobs in his campaign against Sherrod Brown. Fairshake spent $10 million opposing Katie Porter and $2 million against Jamaal Bowman, while Protect Progress dropped $10 million each in support of Elissa Slotkin and Ruben Gallego.

Clearly Carcaise has leveled up from milking long-shot House races.

Segment Insight

This brings us back to Crockett. While doing this little dive, I noticed another mystery vendor in her filings: Segment Insight LLC. Hidden behind a registered agent in Delaware, the company was formed in April 2023. Again, no web presence of note. I can’t claim anything for certain about Segment Insight, but its list of clients looks awfully familiar:

1. Jasmine for Us

• Candidate: Jasmine Crockett (Texas)

• Total Amount: $159,387.00

• Number of Payments: 14

• Purpose: Advertising, Digital Consulting, Email Services, List Acquisition

• Timeline: 2024-02-28 to 2025-06-12

2. Anna Kaplan for New York

• Candidate: Anna Kaplan (New York)

• Total Amount: $117,101.00

• Number of Payments: 4

• Purpose: List Acquisition

• Timeline: 2023-08-17 to 2024-01-08

3. Roland Gutierrez for Texas

• Candidate: Roland Gutierrez (Texas)

• Total Amount: $44,223.00

• Number of Payments: 2

• Purpose: List Acquisition

• Timeline: 2023-10-06 to 2024-01-08

4. Youth Save Democracy PAC

• Purpose: A PAC dedicated to mobilizing young voters to elect Democrats to the House.

• Total Amount: $40,266.30

• Number of Payments: 4

• Purpose: List Acquisition

• Timeline: 2023-09-08 to 2023-12-18

5. Robert Garcia for Congress

• Candidate: Robert Garcia (California)

• Total Amount: $10,000.00

• Number of Payments: 1

• Purpose: Digital Advertising

• Timeline: 2025-06-10

6. Committee to Elect Shomari Figures for Congress

• Candidate: Shomari Figures (Alabama)

• Total Amount: $5,000.00

• Number of Payments: 1

• Purpose: List Acquisitions

• Timeline: 2025-07-30

7. Sites for Congress

• Candidate: Jeffrey Sites (Ohio)

• Total Amount: $1,200.00

• Number of Payments: 1

• Purpose: List Acquisition

• Timeline: 2023-12-01

Overall Total: $377,177.30

Six of the seven reported clients also reported paying Blue Chip Strategies. A side business for Kaple and Carcaise perhaps? Hard to say, but too interesting not to include.

Tail Wagging the Dog

So where does this leave us?

The through-line here isn’t really about Crockett, although the crypto connections are certainly eyebrow raising. It’s about what digital fundraising has become. Blue Chip Strategies perfected an all too common formula: find a Democrat willing to position themselves as a villain’s antagonist, run the content machine, and collect a percentage of whatever comes in. Whether the candidate wins is almost beside the point.

Marcus Flowers raised $16.9 million and lost by 30 points. Nearly a third of that—$5 million—flowed through Blue Chip. Most of that was likely pass-through spending on the actual ads, texts, and emails that fuel the fundraising machine. But consultants take their cut on volume, and volume is what the MTG-antagonist strategy delivers. A campaign that raises $17 million and loses is more lucrative for everyone involved than one that raises $4 million and wins.

Crockett’s exponential fundraising growth—from $970K to $3.1 million to $6.5 million and counting—is legitmately impressive. But it may tell us more about her consultants’ sophistication and Crockett’s media savviness than her electoral viability statewide. Given she’s declared, I certainly wish her the best. I’d much prefer Senator Crockett to Senators Cornyn, Hunt, or Paxton.

Republicans say they’re gleeful about Crockett entering the race. Regardless of her chances, they’re probably correct in that she’ll be useful for their own fundraising as well. That’s the attention economy when it comes to politics: it rewards the fight, not the outcome. Everyone eats along the way.

A Clapback Epilogue

This is something so strange I wasn’t sure where to put it, so I’m dropping it at the end. That very first Facebook ad of the Blue Chip era I mentioned to link to a shopify store called the TheClapbackCollection. Here you can make an actblue donation in return for items like a “Bleach Blonde Bad Built Butch Body” shirt, a “Chile” hoodie or Jasmine Crockett is “My Type” paraphernalia. You can see Crockett herself modeling The leading image for this piece is a screengrab from a video modeling the shirt.

But at the very bottom of the page, above the pay-for box for “Jasmine for Us,” there’s a button reading “Political Consultant-Book Now.”

And when you click it, you’re taken to a page apparently offering “Government Relations and lobbying” alongside other political services.

A WhoIs search shows the webpage was registered in May 2024, after she had been elected, by Crockett staffer Ayala Carmen. The email appears to be one she and her staff used while she was in the Texas State House representing the 100th district.

I honestly have no idea what this is. If anyone can explain why a sitting Congresswoman’s campaign merchandise store has a buried button offering government relations and lobbying services, I’m all ears. A legacy page from her state house days? Staffer mistake? Probably nothing? Bizarre.

Hope everyone had a good Thanksgiving!

One of the odder things I can say about my career is that I cost Bill Cassidy roughly $200,000, adjusting for inflation1. It was the culmination of a year’s work of being very very annoying and finally landing my white whale hit of the cycle. Every researcher has a similar story—the oppo you’ve been chasing since you first caught a glimpse of it surfacing in the ocean of documents you obsess over. Your pitch is your harpoon. Sharpen, reload, and re-fire until you get it to stick in the monster’s flanks.

In 2014, I was Research Director for Senator Mary Landrieu, facing off against Congressman Bill Cassidy. The whale was this: Cassidy was collecting a taxpayer salary for work he wasn’t doing.

Cassidy had been collecting $20,000 per year from Louisiana State University since being elected to Congress, slightly below the cap on outside income for members. Before DC, he’d been a practicing physician and faculty member at LSU’s Health Sciences Center. In his ads and nearly every appearance, he emphasized he’d continued his calling even after being elected to the House.

His financial disclosures claimed this salary merely covered his expenses for teaching. That didn’t sit right with me. Thanks to my brilliant counterpart at the DSCC, we sent records requests to every public entity Cassidy had touched, including LSU. When the response came back months later, we learned LSU agreed to continue paying him 20% of his base $100,000 salary (on total compensation north of $300,000) in return for Cassidy working one day per week. Nothing about expenses. Interesting!

We even got emails from Cassidy himself claiming he’d keep a “paper log” and “a folder in case there are questions.” An LSU official fretted: “We are going to really have to spell out exactly what it is he does for us for his remuneration” because “this scenario will be a very auditable item.” We immediately sent another records request explicitly invoking those words, asking for time sheets, logs, anything. More waiting.

The campaign bumbled on as we began to sink in what was clearly a red wave. By summer, we had our response: LSU could only provide seven months of time sheets. On those, Cassidy reported working 7.5 to 15 hours a month. Not the 7.5 hours per week he’d agreed to! Cross-checking with the congressional record showed half the days he listed overlapped with him voting in DC or on official travel. Incomplete, but enough to show he wasn’t meeting his commitments.

The Cassidy campaign was ready. Every pitch got a deflection: This was LSU’s fault, he’s an at-will employee. Actually the salary just covered expenses. Actually just malpractice insurance. But his own correspondence with the House Ethics Committee contradicted him, LSU provided insurance on top of the salary, he had tenure, and on and on. Still didn’t stick.

We eventually got 16 total time sheets out of 63 months. The salary never changed despite wildly varying reported hours. Some he signed, some he didn’t; three sheets even literally said he “phoned in.” I kept refining the pitch but could never land the story I wanted. One particularly humid night I sat in a bar with an investigative reporter from local TV. Slicked back hair, double shots. I walked him through the hit line by line. He loved it. It never made the news.

We had already failed to win the first round and knew we were going to lose, but at that point I just wanted to bloody his nose. I’d harassed every reporter who would listen and gotten nothing but credulous evasions. It was time to get creative. We handed everything to local bloggers on the promise they’d post the records themselves. Once they were out there for everyone to see, we called a press conference with blown-up copies of his time sheets. The Senator demanded Cassidy prove he wasn’t stealing from taxpayers. If traditional outlets wouldn’t write the story, they’d definitely write about the accusation. We put it in ads, hammered it in our last debate.

The hit was finally in the air.

The Cassidy team scoffed. LSU announced an audit. We lost the runoff. (Though hey, we did two points better than in the first round!)

Months later, after I’d long since left Louisiana, LSU unveiled their findings. They admitted Cassidy failed to “adequately” document his work and roughly 75% of his time sheets were missing or never submitted. But based on testimony from a colleague claiming she saw his white boards full, they determined Cassidy “provided services equal to at least that of his compensation.” LSU found that LSU did nothing wrong. The Chancellor even said they wanted to keep him on payroll.

And yet: upon entering the Senate, Cassidy stopped collecting that salary. His last reported payment was a little under $7,000 in 2014. From 2015 through 2022, he reported no outside income at all. Added up and adjusting for inflation, that’s over $200,000 in lost income.

(That is until 2023 and 2024, when suddenly a new LSU salary appeared on his disclosure. Right in time for his upcoming re-election! Hopefully he’s kept better records this time.)

I’d like to think we made it too politically risky. That counts for something, right?

Except here’s what was actually working while I played Ahab with time sheets and FOIA requests:

“Mary Landrieu voted with Obama 97% of the time.”

That was it. That was effectively Cassidy’s entire campaign. Easy to understand, easier to remember. It came from a website that tracked which bills the White House indicated an opinion on, less than one day’s work to find. In Louisiana, Obama’s approval rating was somewhere between “disastrous” and “catastrophic.” Landrieu was a Democrat. Obama was a Democrat. Voters filled in the gap themselves.

I tried to counter with my own gimmick statistic: Cassidy voted for 97% of the bills Obama signed in the 113th Congress. Which was true! And also completely stupid and ineffective.

Our own polling told a similar story. Our strongest attacks on Cassidy? His support for the Republican Study Committee budget. His vote to raise the Social Security retirement age. His support for cutting Medicare and Medicaid. Nothing complex. No document diving required. Just: he voted to raise the retirement age and cut benefits. One sentence. Everyone understood it.

I remember being put on the phone with James Carville one afternoon. Louisiana native, Clinton war room legend. Basically a pundit by then but a desperate campaign will turn to anywhere for advice. He told me I needed “ice in your veins” and to clobber Cassidy on Social Security. Make it hurt.

“James,” I said, “we’re already running that on TV. It’s been up for weeks.”

There was a pause.

“Well,” he said, “keep it up.”

We were doing exactly what you’re supposed to do, hitting him on unpopular positions with clear stakes. We had the votes, the pain they would cause, everything documented. It just didn’t matter as much as voting 97% with Obama. Sometimes you’re swimming against a tide that is too strong.

I’d seen this before. On Obama 2012, I worked on the Bain Capital team. I had a passion hit about how Bain Capital investments restructured their pensions into “cash-balance” plans. Complex stuff, but the upshot is the transition often ending up hurting the oldest employees the most while saving the employer money. I thought it was a perfect example of how Romney’s policies would harm near or current retirees. Extremely satisfying research. Never got it written, didn’t need to bother. Voters didn’t need financial analysis, they just needed to hear “corporate raider,” “Swiss bank account,” and “shipping jobs overseas.”

Even perhaps the most famous hit of the cycle, the “47%” clip, just appeared online one day. A Mother Jones reporter was handed a video from someone who stuck a camera phone in a fundraiser. No sophisticated research required, no campaign involvement needed at all, and no explanation needed for voters already skeptical of the private equity executive turned politician.

The best hits are the ones that reach escape velocity. They’re straightforward to remember, clear enough to repeat, and they connect to something voters are already inclined to believe. All the detective work in the world doesn’t matter if no one will care.

The art of opposition research is providing vocabulary for what voters feel but cannot yet express. Voters in Louisiana didn’t like Obama. They didn’t trust Democrats. “97% with Obama” gave them language for the idea that Landrieu wasn’t really on their side. It was validation to vote against her.

I still believe my Cassidy research proved he was billing taxpayers for work he didn’t do. I had LSU officials on record worrying about audits. I had his own emails promising documentation he never provided. I eventually got an official investigation admitting 75% of his time sheets were missing. None of it mattered. I couldn’t get reporters to write it. And even if I had, would ‘inadequately documenting work hours at a public university’ connect to any feeling Louisiana voters had about Bill Cassidy?

The more complex a story, the more opportunities to explain it away. Even if you can get it written up, motivated reasoning is powerful. Persuasion is hard. If a voter doesn’t already want to believe something about a candidate, you’re asking them to follow a complicated argument. Most won’t.

There are always exceptions, but I think it’s a good rule for researchers. It’s made me come around on the humble campaign finance hit, where you check who donated to whom and what they voted for. Very little art there. And the arrow of causation is often wrong: politicians get contributions because of what they support, rather than supporting things in return for contributions. But everyone believes politicians are corrupt and bought off, so why not embrace it? It’s much easier to fit in a post, ad, or infographic. Sometimes you might even be right.

It’s also why issue positions and votes can be so powerful. Everyone understands ‘he voted for X’ or ‘she voted against Y.’ Or ‘they agreed’ or ‘disagreed with their party on Z.’ These are very simple cues for voters to recognize. There’s no complexity to deflect, you either supported it or you didn’t. The candidate has to survive the most uncharitable explanation, and that’s on them.

And when conditions are right, simple attacks can overcome even partisan lean. Republican Larry Hogan won Maryland’s governorship in 2014 by repackaging a ‘storm management fee’ into a ‘rain tax,’ distilling everything voters disliked about Democrats into two words. The next year in Louisiana, John Bel Edwards revived David Vitter’s 2007 prostitution scandal, contrasting it with his own record as a Ranger School graduate and West Point alum. Both benefited from unpopular incumbents and angry electorates. But their messages worked because they were simple.

I love playing detective. There’s nothing as satisfying as finding something someone tried to hide, building the case brick by brick, finally forcing it into the open. The mystique is real and I understand why people are drawn to it. Part of me still wonders if we’d gotten those records out earlier, maybe it would have mattered more. But I know that’s almost certainly wrong.

Ultimately the best oppo isn’t the fanciest detective work. It’s finding the words that let the voters say what they want to think. Sometimes that means sophisticated research that becomes a scandal your opponent can’t shake . Sometimes it’s just a statistic on a website.

The ragpicker sorts through garbage to find what someone else can actually use. I cost Bill Cassidy $200,000 with a year of digging and pitching. His campaign beat us by looking up a number on a website.2

That’s the job.

The most complicated phone call of my life took place one evening in Louisiana when I found myself performing tech support for someone with an impenetrable Cajun accent. That person had made a regrettable posting choice on a forgotten account from their past exploration into the performing arts. They did not remember their password, let alone possess a strong baseline of modern computer skills, and I was walking them through how to get back into their profile in order to delete the offending information. I doubt anyone would have gone looking for this youthful indiscretion, but I suppose the long hours helping them clean up a mistake was a mitzvah.

Graham Platner reminded us this week that any online oops is never truly gone. Even if you delete the evidence, as he or his campaign did, it can still resurface. As the oppo guy in a lot of people’s lives, I immediately got a lot of questions about how this could happen. Since I do know a little about this, I can offer how I would go about this and what it says about oppo in 2025.

Andrew Kaczynski graciously shed some light on his process, so we can start there. Platner had linked to his reddit from a personal twitter account he had failed to delete until CNN published their story. That means someone was looking for anything Platner had posted online. They probably already have his email, address, phone number, all the boring identifying information we give to everyone and readily available for purchase or acquisition. You can use those to search for linked accounts, and maybe he made it even easier by posting under his name.

We know from his Reddit posts, Platner was very comfortable sharing personal details online. As shown in CNN’s reporting, even a cursory review of P-Hustle’s posts match with Platner’s bio: soldier deployed to Iraq, former bartender at DC’s Tune Inn, oyster farmer in Maine, etc. It would not be surprising if that was true of the Twitter account as well. Although obviously he didn’t also say anything too interesting on it, given CNN had access to it and didn’t include anything in their story.

But even if that account was a single empty post with nothing but a link to a Reddit comment, that’s all you need to get cooking. Is the comment deleted? Take the link and plug it into the wayback machine or a service like archive.is. Click on whatever handles you see in the linked thread. Their profile pages , along with their history of posts, have probably been crawled by different archival bots. Which is exactly the case in CNN’s story, which links to the archive.is copy of P-Hustle’s profile page. From there we see posts clearly marking the account as Graham Platner, confirmed by him under direct questioning.

The Washington Post even linked to a fancier tool, PullPush, which is functionally a searchable mirror of historic Reddit data. Starting in 2015, Reddit partnered with the social media collection platform Pushsift which used Reddit’s own API to scrape data from the website within a few hours of posting. This data was made freely available to academic and other researchers until 2023, when access to Reddit’s API with tightened dramatically and only approved Reddit moderators could use Pushshift.

Historical collections of the older datasets still do remain available via tools like PullPush, which can be queried and search using all kinds of filters like author, subreddit, etc. PullPush will generate JSON files containing all posts falling under whatever parameters you set. Out of professional courtesy I won’t literally walk through how to download every post made by Platner prior to 2023 for review. But these days it’s as simple as reading the documentation and using an llm. And though archival web crawlers have been restricted by modern reddit, the period when Platner was posting heavily was well covered by tools easily available to anyone.

These methods are indistinguishable from investigative reporting, or OSINT if you’re feeling fancy. Why you are using these tools is what makes it opposition research. Why are you interested in Platner? Because you want to imperil his candidacy. The lowest hanging fruit to doing so is finding him saying something dumb. You know that if someone is a poster, they have probably posted something dumb. So if you want to make someone look bad, you should look at anything they have posted online. Easy to understand, easy to do, no Spotlight style shoe leather needed.

Does that mean this story was a pitch? There is some mitigating context. KFile at CNN operates similar to any old fashioned oppo shop with a full team of a talented reporters like Em Steck. Andrew Kaczynski even got his start as an intern at the RNC years and years ago. I first came across him while working on the opposition research team for Obama’s 2012 re-elect in Chicago. We were tasked with examining the menagerie of potential Republican candidates, a “target rich environment” in President Obama’s words when he stopped by our table on one of his rare visits to the office. At this time I had come across C-SPAN’s then new and expansive online video archive, helpfully tagging appearances dating back decades. Fantastic resource for old clips of Romney defending Roe v. Wade, Newt Gingrich supporting an individual mandate, Rick Perry talking about sausages, and Herman Cain being a weird delight.

You’ll notice I posted links to these examples from Kaczynski’s youtube rather than C-Span. Well suddenly these videos were showing up online and in stories. I was even asked directly if was I leaking them myself rather than allowing our communication leadership decide how and where to deploy research for pitching. But the true answer was a then undergraduate Kaczynski had found the same clips as me and was posting them just for the fun of it. Ben Smith rightly brought him into Buzzfeed and the legend of KFile was born.

So while CNN certainly accepts pitches and researchers are digging into Platner, it’s reasonable to believe this story was not purely the product of some nefarious oppo dump. As a high profile candidate with an unorthodox background, of course a reporting team that specializes in the art of internet necromancy would kick the tires on Platner. It doesn’t really matter, as the outcome is the same. And once Platner confirmed all the details in CNN’s story, it became trivial for other journalists or oppo flacks to recreate their methods.

These follow up stories likely came from a mix of journalists following KFile’s path and from opponents pitching additional material. You can squint and possibly see a hint of this in Politico’s piece, which devotes four paragraphs to emphasize the similarity between Platner’s posts and violent text messages sent by the Democratic nominee for Attorney General in Virginia. To me that reads like the thematic framing one would include in a “hit,” where the research is shared alongside additional information like the details of the Virginia case and how that’s hurt their candidacy. The hope is the reporter sees this as an easy to write up 800 words and adopts your preferred frame: basic public relation stuff, just applied to oppo.

We can’t know for sure, and there’s nothing wrong with that even if it is the case. Oppo flacking is honest work and the news has to be made somehow!

But why wasn’t Platner’s team prepared for this story? My working assumption, based on my past whiffs here, is that Platner himself knew he had a reddit problem. He or his team deleted the account. But nobody caught the old twitter account linking to it because they forgot it existed or the vet just missed it. These kind of things happen more than anyone cares to admit. I once let a guy who violated the embargo on Iran on stage to speak at an event with my candidate. Whoops, it happens. Tightened the traps after that one, which thankfully no one caught. I’m becoming reliable in my defenses of researchers but you should never expect perfect from anyone.

If his advisors knew the full extent of his comments, should they have stopped his candidacy? That gets the relationship between candidate and consultant backwards. Advisors advise, the candidate decides. Former Mississippi Governor and Super Lobbyist Haley Barbour famously paid for a research report prior to his potential candidacy in 2012 and decided a campaign wasn’t going to be worth the headache. In Platner’s case, he clearly came to the opposite conclusion. And if your candidate is in, so are you; especially if you think they can raise the money to give you a better than zero chance of running through the tape. (Granting exceptions for particularly vile or potentially illegal behavior).

If I knew then what we know now, and that’s the full extent of it, I wouldn’t have thrown my body in front of a run. The ‘communist’ line will echo forever, but you can also see he’s sincere about the second amendment and losing faith in Collins - both plusses in Maine. The voters have repeatedly proved immensely forgiving of racism and sexism from candidates, so maybe they’ll extend that grace to edgy shitposting.

Platner has millions in the bank and enough time to wait for the polls. It’s unsatisfying, perhaps, but there’s something refreshing about deferring major life decisions to statistical sampling. If it’s truly fatal, he’ll know in the coming weeks. But typically no one suffers much for trying to stick it out beyond a bruised ego and the apology tour circuit.

The ultimate lesson here is that nothing you write online ever truly disappears. I do internet archaeology for a living and I can tell you all those jokes you deleted in 2015 are still out there. The only question is whether anyone bothers to look. Usually they don’t. But if you run for office, someone always will. Whether by way of oppo researcher or investigative reporter or some combination of the two, negative information wants to be found and people want to find it. Enough people looking for anything on anyone will eventually discover something.

(Photo by Tim Mossholder on Unsplash)

On October 2nd, the Trump administration quietly warned their mass deportation campaign is creating a ‘structural crisis’ on American farms that threaten the nation’s food supply. Their solution? Import 119,000 mostly Mexican guest workers, while paying them less than what they estimate unauthorized workers currently earn.

This admission came in the form of an Interim Final Rule published in the Federal Register. The American Prospect and Washington Post have both ably covered this story and I particularly recommend David Dayen’s piece for highlighting the broader contradictions at the heart of the rule. I’ve been obsessing over this rule for weeks, and the wage math is even worse than the headlines suggest.

In the interest of full disclosure, I didn’t pitch these stories. Smart reporters simply noticed the same thing I did. But now that they’ve covered the important stuff, I can dive into the granular details that no editor would let see print.

I’m going to cite Philip Martin extensively throughout this piece. He’s the leading expert on immigration and agriculture, cited in Trump’s new rule and featured in John Oliver’s investigation on farm labor abuses. Consider this a curation of his research.

The Farm Workforce: Mostly Mexican But Changing Status

Let’s start with an overview of the American agricultural workforce. Philip Martin estimates the 115,000 US farm establishments employ about 2.5 million workers annually, 80% of whom were born in Mexico. That number is divided between 2 million crop workers and 500,000 animal agriculture workers. 850,000 of these workers are unauthorized while the remaining 1.6 million have legal status. That includes roughly 315,000 H-2A seasonal visa recipients, 92% of whom come from Mexico.

Most of these unauthorized workers entered during years of lax enforcement in the 1990s and 2000s and are now in their 40s and 50s. This aging workforce is less productive and less able to do strenuous manual labor. But they’re not being replaced with new unauthorized workers. Stricter enforcement in the mid-2000s, the 2008 recession, and improving conditions in Mexico all reduced migration flows into the US.

Martin’s analysis of Department of Labor surveys shows this change clearly: the unauthorized share of crop workers started at 10% in 1989-1990 and peaks 55% in 1999-2000. By 2013-2014 it had fallen to 47% and now sits at somewhere around 42% of crop workers. This decline began long before Trump took office and has been ongoing for decades.

That demographic shift has a practical impact: the workforce has become far less mobile. As workers age and settle in one location, far fewer qualify as “migrants” who travel between seasonal jobs. That creates a labor availability problems for farmers since crop seasons vary by location and crop type. Even the recent surges in unauthorized migration under Biden didn’t solve this, as new migrants overwhelmingly preferred stable city jobs over demanding seasonal farm work.

The Rise of the H-2A Visa

In response, employers have turned to another option: the H-2A visa. H-2A workers are usually younger, better educated, and more productive than older workers of any immigration status. But they’re also expensive: Martin estimates H-2A workers cost $30+ per hour in California versus $20+ for U.S. farm workers, once you factor in recruitment fees, processing costs, housing, and transportation.

The Biden State Department produced a nice graphic showing how the H-2A visa has exploded over the past decade: from less than 50,000 in 2005 to over 300,000 today.

The new rule introduced on October 2nd concerns the “Adverse Effects Rate Methodology” or AEWR. The AEWR has existed in some form since the 50s to stop guest workers from undercutting American wages and employment. The Eisenhower Department of Labor found farms that primarily employed guest workers paid less than farms that employed only US workers. So it instituted state level mandatory wage floors for employers of guest workers, applying both to those workers and any U.S. employees in ‘corresponding’ employment.

Today federal law requires employers to pay H-2A and corresponding US workers the highest of the AEWR, state minimum wage, federal minimum wage, or collectively bargained wage. The AEWR is usually on the high end of that range. It’s set at different rates for 15 farm labor regions, with individual rates for California, Florida, and Hawaii based off surveys of farm workers wages. Employers usually argue the methodology is wrong and pushes wages too high, while worker advocates say it’s too low and is frequently violated.

The Trump Department of Labor is unsurprisingly siding with employers, proposing to replace the AEWR with a brand new methodology. Their justification rests on three core claims. Let’s examine each one.

Claim 1: Deportations Raise Labor Costs Due to the AEWR

In their words, the “unreasonably high FLS-based AEWRs were only workable because agricultural employers could turn to low-priced illegal aliens,” which is now under threat due to Trump’s deportation campaign.

“Given the scale, speed, and investment in the federal government’s efforts to enforce immigration laws and restore the integrity of the U.S. border, the Department concludes that there will be significant labor market effects in the agricultural sector, which has long been pushed to depend on a workforce with a high proportion of illegal aliens. Because these illegal aliens often possess specialized skills suited to agricultural tasks and typically earn lower wages than authorized workers, their sudden and large-scale departure is expected to significantly increase labor costs for employers. These cost increases are very likely to limit the ability of agricultural operations to maintain current production levels or expand employment, resulting in downstream impacts on food supply and pricing.”

The timeline here is questionable at best: instead of H-2A forcing farmers to depend on unauthorized workers, in reality the long term decline of unauthorized workers is what has pushed farmers to the H-2A. These trends predate both Trump administrations.

Claim 2: Higher Labor Costs Put The Food Supply At Risk

But now, according to the Department, “without prompt action, agricultural employers will face severe labor shortages, resulting in disruption to food production, higher prices, and reduced access for U.S. consumers, particularly to fresh fruit and vegetables.”

In their words, there “is ample data showing immediate dangers to the American food supply.” They describe a “structural, not cyclical, workforce crisis”—driven by the loss of what they call the “mobile illegal alien workforce.” Citing the decline in illegal border crossing and increased enforcement, they claim this “near total cessation of the inflow of illegal aliens” threatens “the stability of domestic food production and prices for U.S consumers.”

Ominously, they even suggest “[s]uch significant economic impacts not only create tangible and imminent economic harms, but they structurally disrupt the ordinary operations of the U.S. agricultural sector, resulting in shortages of agricultural commodities that cannot be supplemented with imports in the near-term.”

Sounds bad!

Claim 3: American Workers Are Simply Too Expensive

The following paragraph sums up their entire argument. Farmers are “economically reliant” on low paid unauthorized workers and even higher wages aren’t enough to tempt Americans into farm work:

“Decline in the illegal alien population will only exacerbate this already pressing mismatch in the agricultural labor market and deprive growers of a relatively cheaper labor supply on which they have become economically reliant. (A substantial body of research estimates that illegal alien workers earn between four percent and 24 percent less than similarly situated legal workers, giving employers a strong financial incentive to hire illegal labor.)  Despite rising wages, there is no indication that unemployed or marginally attached U.S. workers are entering the agricultural labor force in meaningful numbers. Without swift action, agricultural employers will be unable to maintain operations, and the nation’s food supply will be at risk.”

Keep that statistic of unauthorized workers earning 4%-24% less than similarly situated authorized workers in mind for later. That figure is the academically accepted ‘wage penalty’ faced by unauthorized workers once you adjust for factors like age, education, and other demographic characteristics. The Biden White House linked to several of these studies here if you’re interested.

Back to the rule. Here’s where it gets truly revealing. The Department doesn’t just say wages can’t go up—they say American workers are being unreasonable to expect them to:

“Turning to the potential reliance interest of U.S. workers in the current methodology, the evidence relied on throughout this IFR strongly indicates that such reliance is tethered to a labor market that is dramatically changing and increasingly unstable. As discussed, the current and imminent labor shortage and the subsequent natural correction of a labor market artificially impacted by illegal aliens cannot be avoided. The Department simply has no evidence of the existence of a substantial population of U.S. workers who are willing and able to accept wage rates that are reasonable and proportionate to agricultural work but are deterred from entering agricultural work by AEWR-priced H-2A workers.”

In other words, any wages that an American would accept are in turn unacceptable to employers.

The Solution: Cutting Wages for EVERYONE

This new rule replaces the old AEWR with a new two-tier skill level system for every state. The Administration estimates at least 92% of guest workers would fall under the entry level “Skill Level I,” as 92% of H-2A workers are now paid at the base AEWR rate.

Remember the claim unauthorized workers earn 4-24% less than legal workers? According to my calculations, the cut from current AEWR to Skill Level 1 would reduce legal wage levels by 8.3% to 35.9%. So wages for H-2A and corresponding US workers could now be set below what the administration estimated unauthorized workers presently earn!

Here’s my analysis breaking down the wage cuts by state:

But it gets worse, as the rule goes even further to cut wages. I mentioned earlier that H-2A workers are often more expensive once you account for employer provided housing and similar non wage compensation. The Department declares this deeply unfair for American workers who pay their own rent and requires an additional pay cut:

“the Department is implementing a standard downward adjustment to the hourly AEWRs that accounts for the compensation disparity U.S. workers face when H-2A workers are being paid for work performed under the same work contract but, unlike most U.S. workers, receive additional non-wage compensation in the form of free housing. Those U.S. workers who are reasonably able to return to their permanent places of residence at the end of each workday, must continue to bear these essential costs from their wages, despite often being offered and often paid the same wages as H-2A workers. Thus, the result is an adverse disparity in compensation where the effective wage rate of U.S. workers is lower than that mandated for H-2A workers under the same work contract, which the Department views as prohibited by the statute that this IFR seeks to correct.”

The traditional economic logic would be if guest workers cost more, employers have an incentive to hire cheaper U.S. workers instead. The Department flips this: they claim that cutting H-2A wages protects American workers by eliminating the ‘compensation disparity.’

Once you include this housing adjustment, which only applies to H-2A workers, they can now be paid anywhere from 15.5% to 44.4% less than before.

The Upshot: Lower Wages for All and More Guest Workers

To be clear, employers must still pay H-2A workers the highest of the adjusted AEWR, prevailing, collectively bargained, and minimum wage rates. Still the Department of Labor estimates this rule change will serve as an over $20 billion wage transfer from H-2A workers to their employers.

The Department even admits that the AEWR cut “also represents a wage transfer from corresponding workers, not only H-2A workers” but “the Department lacks sufficient information about the number of corresponding workers or their wage structures to measure these impacts.”

Translation: it’s a wage transfer from American workers whom either must accept lower wages or be replaced by cheaper guest workers.

The Department expects this rule to bring in 119,000 additional guest workers by 2034, from a projected employment of 383,000 to 502,000. But as Dayen notes, some farmers are worried they may not be able to attract workers at these lower wages. Meanwhile Washington Post quoted immigration restrictionists have criticized this about-face on prioritizing American workers. Project 2025 even proposed phasing out H-2A altogether in favor of subsidizing the mechanization of agricultural labor.

Mechanization would take years and billions in investment. This administration has chosen the faster path: a largely invisible expanded guest worker program with limited labor protections and potentially lower wages than the unauthorized workers they’re deporting.

The argument is cynical, yet elegant in its own way: Deport unauthorized workers to create a ‘crisis,’ ignoring that the demographic shift in question has been happening for decades. Use that crisis to justify importing more guest workers than ever. Slash their wages below what you said unauthorized workers earned, undercutting American workers in the process.

And call it ‘America First.’