Defector Annual Report, September 2023 – August 2024

Published on October 16, 2024

Purpose of this report

This is the fourth year that Defector’s business team is publishing an annual report. Let us be frank: We sort of regret backing ourselves into this corner of having to produce one of these every year. We’re not sure we will continue to have all that much interesting to say, and writing clear, digestible prose is hard. (In some ways the great value of this exercise is to be reminded of how difficult our coworkers’ jobs are. How does anyone blog day after day?)

This year, we are going to streamline the report. After the Financial Summary, which will once again provide a year-over-year view of how Defector’s money was earned and spent, we will highlight a handful of topics that our subscribers might find informative, interested media parties might find useful, and subject matter experts might be able to help us with.

For a more comprehensive picture of the inner operations at Defector, go ahead and refer to our annual reports from Year 1, Year 2, and Year 3.

Financial summary*

The question of distribution

With the AI-fication of search and the ongoing degradation of Twitter, digital media participants and observers have all been wondering: Where should publications look for distribution? (Internally we define distribution in the broadest terms—getting Defector blogs in front of interested readers, subscribers and non-subscribers alike—though for the purposes of this section, we are really focusing on new readers with zero or light familiarity with Defector.)

Finding a satisfying answer to that question is perhaps less urgent for Defector than some other media companies, as the bulk of our revenue continues to come from subscriptions. Downturns in search and social traffic therefore have much less direct impact on our near-in income statement when compared to publications reliant on display advertising or affiliate revenue. But it’s still an existential question in the long term. Subscription businesses are leaky buckets, where even the best businesses lose some percentage of subscribers month after month, and we must find ways to refill the top of the bucket.

In Year 4, subscriber churn was low and retention was strong, but new subscriber acquisition most months was slower than forecasted. Our Year 4 high-water mark of 42.5k net active subscribers was a single percentage point above the Year 3 high-water mark of 42.1k. Within the context of sluggish new subscriber acquisition, the question of distribution looms large.

Not surprisingly, there do not appear to be any simple, one-size-fits-all answers. Some ideas that we’ve pursued, to varying levels of success and scalability:

Gift links: In addition to making a couple dozen less-overt site improvements this year, our pals over at Lede by Alley launched two of our subscribers’ most-requested features: Dark Mode and gift links. Since April, Defector subscribers have had the ability to generate paywall-bypassing gift links which make it easier to share blogs with friends, family, and group chats. Our goal is to (partially) offset the loss of many, middling-quality referrals via social media with a smaller number of high-quality referrals via gift link.

The early stages of the program are encouraging, and we’ve learned a couple things already. First, and this advice is probably true of any launch, is that you really gotta let people know that a feature has gone live. No matter how obvious you think the new Gift This Article button is, many people will simply never notice it. Despite gift links being live on the site for six months now, we still get occasional emails requesting the feature.

Second, we’ve had success drawing attention to the feature by running a couple “Free Fridays”—dedicated weekends on which all of our social media posts and newsletter links contain gift links. There’s a real, tangible cost associated with doing this, because email signups slow down significantly when visitors aren’t prompted to register. However, Free Fridays seem to boost both new visitor pageviews over the weekends on which they run, as well as subscriber gift link generation in subsequent weeks (2–3x the pre-intervention rate). We are happy to make this trade occasionally.

Finally, gift link usage has the “shape” we hoped for—a handful of links have gone moderately viral on social media, but an overwhelming majority of gift links are viewed by just one recipient. This looks exactly like the high-quality referral to a friend use-case we wanted to encourage. Despite not having a monthly limit on gift link generation, we have also seen zero signs of abuse of the feature. Thank you to each of the tens of thousands of subscribers who have (responsibly) shared gift links!

Newsletter ad buys: It makes intuitive sense (and is substantiated by survey data) that advertising Defector in likeminded, subscription-based publications would be a cost-efficient way to acquire subscribers. Unfortunately, our experience does not yet support this intuition. During Q2, we spent a few thousand dollars to take out sponsorship units in a handful of newsletters we perceived to be read by “Defector-adjacent” people. In total, these ads generated a few dozen subscriptions. Assuming normal retention rates among these subscribers, the ads ought to eventually pay for themselves.

However, focusing solely on customer acquisition cost and payback periods obscures a bunch of other opportunity costs involved with advertising in this way. For one, it’s hard to write decent ad copy. We would likely get better and more efficient at this with additional practice, but it’s time-intensive to match each newsletter’s tone and curate blogs for its readers. Even more time-intensive is the process of pitching and coming to terms with the operators of prospective newsletters. Many of the people to whom we reached out don’t regularly run ads and therefore didn’t have a standard ad unit offering, a going rate for that offering, or even a means of taking payment. As people who’ve historically done a lackluster job at selling our own ad units, this makes perfect sense to us: The possibility of irregularly receiving a couple hundred dollars hardly seems worth the effort of creating a robust ad sales process.

We nonetheless appreciate all the newsletter operators who were willing to take our money. Thanks especially to the one who ended up being the only person who bought a Defector subscription using their newsletter’s promotional code. Without publicizing anyone’s rate cards, we were also shocked by the large gap in CPO (cost per thousand opens) across some reasonably established newsletters.

Zig-zag regwall strategy: As a site, our goals still boil down to: (1) reaching as many people as possible with our work; and (2) making enough money to keep running Defector on our own terms. These goals are broadly compatible but intermittently conflict with one another. One site of conflict is our registration wall (“regwall”) and paywall rules: Putting blogs behind a gate dampens their reach, but failing to gate blogs reduces the amount of subscription revenue generated by each blog.

For most of Defector’s first three years, we navigated this contradiction by running a dismissible regwall under which logged-out visitors would encounter a dismissible, slide-up prompt to register an email to continue reading the site. The prompt typically became non-dismissible after reading a few posts (i.e., you then had to fork over an email address or pay to read any more blogs that month).

In 2024, we began experimenting with “zig-zagging” between relatively more and less restrictive regwall regimes, for two reasons. First, there’s probably just some utility in being unpredictable. If potential subscribers know that they’ll encounter an impermeable paywall each time they click on a Defector link, they’ll learn to stop clicking on those links. Gift links are also useful in this regard. Second, we expected that some number of anonymous visitors would tend to get in the habit of reading the site and build up during “soft” regwall periods. Briefly cutting over to a “hard” regwall would allow us to capture emails from these readers; subsequently reverting to a soft wall would allow us to resume accumulating new ones.

This strategy has seemed to work: Making the regwall less restrictive (by eliminating the dismissible slide-up prompt on a visitor’s first few article reads) tends to reduce email signups by about a third. It also allows visitors to read more blogs. On the other hand, tightening the regwall (by requiring logged-out visitors to register an email address in order to read any blogs at all) tends to roughly triple the rate of email signups.

We’re excited that a month consisting of three weeks of loose regwall and one week of tight regwall tends to generate ~25% more email signups than a month of steady-state dismissible regwall. Generating email registrations such that we can send newsletters and subscription appeals directly to potentially interested readers is the best available hedge against our degraded and degrading reach on social media.

Apple News: The Apple Publisher Relations team first reached out to Defector in late 2021 to gauge our interest in publishing to the Apple News app. There were two options on the table. One, we could publish a few one-off Defector blogs each month, manually selected in conjunction with the Apple News Sports editors. There would be no revenue share agreement; Apple would get a couple extra posts in the Sports category for their free Apple News product, and Defector’s name would get in front of a different universe of readers, with calls-to-action to give us their email addresses. Or two, we could participate in the Apple News+ program, publish every Defector blog to their app, and get a share of Apple News+ subscriber revenue.

After speaking with people at a few publications that had recently faced the same decision, we could not comfortably answer the critical question on whether the Apple News+ path would meaningfully cannibalize our existing subscribers and most-likely-to-be-converted readers. We decided to go with the first option, with the vague promise to revisit the question later as Apple and other publishers got more data. We then forgot to revisit the question until Semafor’s report about Apple News+ delivering significant revenue to other publications. At the moment, Apple News+ is not accepting new publications, so we’ll kick the can down the road on whether it would be in Defector’s long-term interest to participate in the Apple News+ program.

Serving ads to non-subscribers

Later this year, Defector.com will begin regularly serving display ads to all non-subscribers. BuySellAds, who sell ads for our independent media compatriots at 404 Media, will be our exclusive direct ad broker.

Why are we focusing meaningful attention on the ads business for the first time, in the fifth year of Defector’s existence? Well, we’re hoping to do more hiring across a couple of disciplines in the near-future, and a frank accounting of the current state of our business versus our goals for the company leads to a pretty obvious truth: A subscription revenue stream expected to grow in the low single-digits cannot on its own fund the additional hiring we’d like to do. There are generally two ways to make money publishing words on the internet, and we’ve largely ignored one of them so far. 

We’ve previously sold on-site advertising, in fits and starts, presented to subscribers and non-subscribers alike. Warby Parker sponsored our launch month in 2020, and we’ve had sponsorships from a small assortment of direct-to-consumer brands over the ensuing years, largely through inbound interest. But on-site advertising never became a repeatable, reliable source of revenue in our first four years, largely because of the site’s non-standard advertising units and the reality that our operations team members are bad at sales.

Having more regular display ads served to non-subscribers should strengthen the value of our paid subscription offerings, as our paid subscribers will see the contributions of our eventual new hires while getting an ad-free experience that they’ve largely come to expect previously (though you’ll likely start to see some ads in the weekly Blogs of the Week newsletter). The Lede team is working on the technical details to run ads only for non-subscribers, and we’ll extensively test that functionality before going live with BuySellAds; some subscribers may hear from us to help with early testing. In the past, certain subscribers have written to us explicitly stating that they would gladly view and click on display ads as another means of supporting Defector. God bless you, that’s a wild thing to volunteer. Lede may build in a toggle for “show me ads” in your subscription Account Settings in the future, but that will not be part of the initial rollout.

As part of our internal discussions before moving forward with BuySellAds, we’ve carefully considered our relationship to our non-paying audience, and what we owe to them and what they owe to us. Every month we have a seven-digit number of unique visitors, but our database of registered emails stays in the range of 200-300,000, and our paid subscriber count hovers a bit above 40,000. Ideally, more and more of those unique visitors would be giving us their email addresses, and those email addresses would eventually convert to paid subscribers, but the reality is that millions of visitors will never consider giving us money directly.

Which is OK! Many of them had never visited Defector.com before and won’t be back, or they read Defector infrequently enough that they never hit our hard paywall, or they have never paid for a subscription for written media, or they can’t afford to give us $8 a month. But then what do we owe each other? Where our staff has landed is: Defector owes them a browsing experience that doesn’t cause their laptops to overheat from a gigantic ad load, but we don’t owe them an ad-free experience. And we think it’s reasonable to ask a visitor who will never give us money directly to look at some number of ads, from high-quality brands, and generate fractions of cents for us. If you’re a regular Defector reader who is unwilling or unable to pay for a subscription, we’d ask that you keep your ad-block turned off to support us.

Defector’s health insurance journey

In the first few years of Defector’s existence, the annual revisiting of “what are we going to do about health insurance” has been perhaps the single most unpleasant, most opaque, and most difficult-to-navigate operational issue. Now that we are on more stable footing for the foreseeable future, it’s worth laying out some details of the health insurance journey.

Defector’s relationship with our outside HR partners at Your Other Half (YOH) began as a one-off consulting project in August 2020. They were contracted to evaluate the possible pathways to acquire group health insurance for our unique company, which as of summer 2020 had nineteen owner-employees across eight states and, uh, zero income (to be fair, we did have over a million dollars in prepaid subscriptions in the bank, but we couldn’t touch that money or recognize it as revenue before we started publishing blogs and providing an actual product).

YOH solicited quotes on buying group insurance directly on the state exchange in New York, where Defector Media LLC is headquartered and the largest number of our staffers live. They also convinced a handful of Professional Employer Organizations to bid on our business. The PEO business model is to provide payroll, benefits, and tax and regulatory compliance services to small and mid-size businesses; in broad strokes, they can aggregate all the employees for whom they are a “co-employer” as one big insurable group and negotiate plan pricing with the major health insurers. The pricing from the New York state exchange was exorbitant, so we were choosing between PEOs. Trinet offered the best all-in pricing across health insurance and payroll administration, so we happily signed on to work with them.

But for several years afterwards, we waited for our annual health insurance rate renewals with bated breath because a PEO can allocate highly variable price increases as they see fit across the small businesses they work with, regardless of what overall price increases they negotiate with the insurers. The PEO can set those price increases based on healthcare service usage the previous year, changes in the demographic makeup of the employee base, or whatever other criteria they want to consider.

We left Trinet and started using Insperity as our PEO in 2022 in pursuit of more manageable health insurance costs, but as of early 2024, our company’s per-employee health insurance spend was nearly 2x what it was 36 months prior.

Coming up on our Spring 2024 renewal, we were stressed. We knew that several staffers had undergone major procedures in the past year, so we were expecting potentially significant rate increases. YOH put our business out to bid again, but they added a new prong to their approach: In addition to getting pricing from PEOs and the New York small business health insurance exchange, they would find direct brokers for the state small business exchanges in every state in which Defector has any employee living and working (and thus can prove we’re legally doing business within the state) and inquire on group insurance pricing.

This approach was extremely fruitful. Not only did we get better pricing, such that for the first time we could expect to save money on health insurance versus the prior year, but state exchanges are required under the ACA to cap year-over-year pricing increases on all plans to 15%, so our annual period of intense anxiety as we await renewal pricing will be much less agonizing going forward. The downside, which we’re only semi-joking about, is that the staffers who live in that state basically can’t move, lest we become ineligible to purchase off that state exchange anymore.

Alongside getting health insurance from a state exchange, Defector now uses Justworks as our PEO for payroll services and dental and eye insurance. We’re very glad to have arrived at this health insurance arrangement, and hopefully we won’t have to revisit it anytime in the near future. But the journey once again speaks to how much it sucks shit to navigate the health insurance system in this country. Groups of journalists approach Defector all the time to ask for advice on how to quit their jobs and start new publications. And a recurring and completely understandable barrier for them to strike out on their own is, can they and their families really afford to give up access to employer-sponsored health insurance? Defector has multiple full-time operations people, professional HR support, and a healthy balance sheet, and it still took us years to land on a workable health insurance arrangement. How are nascent small businesses, media or not, supposed to figure out a sustainable health insurance path (if it even exists) on their own?

How do you change organizational structures within a cooperatively owned and operated business?

The editorial team at Defector recently undertook some meaningful changes in team processes and organization. The details of the evolution don’t much matter; people who work in other newsrooms would find the adjustments straightforward to the point of boring. But interested parties do sometimes ask how we agree on changes in a worker-owned company where every employee has a voice and a vote, and this may be a useful illustrative example.

Although we’re often tinkering with day-to-day processes, over the past year some staffers began to express interest in more significant changes in the blog manufacturing system. As we planned the programming for our team offsite in May 2024, multiple people asked for dedicated meeting time to discuss the editorial system that week when we would all be together in New York.

Jasper often jokes that being “good at business” simply boils down to running effective meetings. An effective meeting requires a clear vision of what has to happen before and after the meeting. What would be the ideal set of next steps coming out of the meeting? Who needs to be involved and what are the necessary analyses, conditions, and inputs that have to happen before the meeting to move everyone towards those next steps?

Working backwards, we checked on who would eventually have ultimate decision authority for implementing major changes to the editorial system. Upon reviewing the company operating agreement and all the decisions outlined in the company RAPID framework, it was not clear where the decision authority lay, though it was clear that such a decision had not been assigned unilaterally to the Editor-in-Chief. 

So, before we held any discussions about any specific proposed changes, we had to do the meta-work of laying out a process for ratifying any changes. We wanted a two-thirds affirmative vote across the editorial team, as a simple-majority vote felt a touch flimsy and might result in a too-sizable unhappy minority. We went with the following process:

1. Every editorial staffer was welcome to propose changes to the existing editorial structure. They were asked to construct memos laying out their visions, with memos to be distributed as a pre-read the week before May Meetings.
2. During May Meetings, we set aside an afternoon dedicated to the memos. Each memo writer would present their vision, and everyone else could ask clarifying questions, make suggestions, and otherwise engage with the proposed ideas.
3. After May Meetings, each memo writer had a few weeks to rewrite their memos to include feedback as they saw fit, and to answer a more detailed list of questions, so that everyone could understand the similarities and differences across proposals on a few key points. Memo writers were encouraged to combine efforts with other writers whose proposals were similar to their own.
4. Once we arrived at a final set of memos, we ran a two-round voting system:
   1. Round One was ranked-choice voting, a la New York City local elections.
   2. The winning proposal from Round One then moved to Round Two, which was a simple up-or-down vote. If the proposal received Yes votes from a two-thirds majority, the vote would pass and it would be adopted by the organization.
   3. If the proposal received less than two-thirds Yes votes, it would fail to pass, and we’d effectively default to having the Editor-in-Chief implement any structural changes to the existing status quo as he saw fit.
   4. Nobody on staff would know how anyone else voted, which required that we ask an outside party to count the ballots and make sure no one accidentally double-voted in the Google Form.

Beyond a couple deadlines having to be extended, the process ran smoothly and resulted in the adoption of a memo. The changes have largely been implemented, though we expect to work through some more growing pains, especially as the items that have to happen monthly, quarterly, and annually occur for the first few times.

Thanks to our freelance contributors

In Year 4 we commissioned freelance pieces from 80+ writers, accepted crossword submissions from 30+ constructors, and worked with a number of producers, illustrators, and photographers. We’re so grateful to all of these contributors for making Defector a more vibrant, interesting website. Below are some particularly excellent freelance blogs that generated an outsized number of pageviews or new registrations:

• The Creator Of 'Magic: The Gathering' Knows Exactly Where It All Went Wrong by Nick Zarzycki
• The Life Of An International Minor League Baseball Player Has Never Been Simple by Jen Ramos Eisen
• 'American Fiction' And The Wet Eyes Of The Sentimentalist by Jason England
• The Judgment Of Magneto by Asher Elbein
• The Life Of That Unhappy Bobby Knight Is Brought To An End by Charles P. Pierce
• Why Your Team Sucks 2024: Buffalo Bills by Victoria Zeller
• The Ghosts Of New Atheism Still Haunt Us by Erik Baker
• At The Olympic Trials Marathon, DFL Means Pain Over Regret by Dennis Young
• The Yamuna River’s Treasure Divers Fight More Than Poverty To Survive by Meher Qadri
• The Doubts Were Not Reasonable by Tom Scocca

Additional odds and ends

The default billing cycle on the product page significantly impacts subscribers’ choices. A monthly Reader subscription ($8/month) notionally costs $1.42 more each month than an annual Reader subscription ($79/year). However, based on Stripe’s cut being >2x as large on 12 monthly transactions than on a single annual one, and the substantially higher churn rates of monthly subscribers, a new annual subscription projects to be worth ~20% more to us during its first year than a new monthly subscription.

We were therefore encouraged to find that making annual subscriptions the default on our product page boosted annual subscription uptake by ~30pp. However, this mix improvement is surely some combination of nudging indifferent subscribers toward an annual option (growing the numerator = good) and scaring away people who might otherwise have bought a monthly subscription (shrinking the denominator = bad).

We don’t have the ability to A/B test this, but we’re sure that untold other businesses have. If you have a sense for the relative impact of those numerator vs. denominator effects, especially in a subscription media context, please email Sean!

Onboarding campaigns work: As we’ve shared in past reports, the count of distinct days on which a subscriber reads something on our site is the single best engagement-based predictor of churn risk. This is pretty intuitive, since subscribers who show up frequently are subscribers for whom reading Defector has become a habit. We continue to find that “onboarding” email campaigns are effective at helping make Defector a habit for new subscribers. A simple, two-email series we instrumented earlier this year appears to have boosted first-month engagement by ~2 days and ~6 articles read. These engagement gains appear to have flowed through into a 5-10pp reduction in first-month churn rates.

Discounted introductory offers also work, and their effectiveness can be amplified by onboarding campaigns and the default billing cycle. We added ~2.5k subscribers through a pair of sales we ran this summer: one for Normal Gossip, the other for our regular Reader product. This year’s regular business sale was responsible for two of our three largest weeks of subscriber acquisition since 2020. As a result of making our default offer an annual subscription, two-thirds of those trial subscriptions were set to transition to annual plans, roughly twice last year’s rate. Nearly 75% of all trial starts have successfully transitioned to paying full price. That number is slightly worse than last year’s 80% survival rate and is likely attributable to the monthly-annual mix shift (i.e., people scheduled to pay an annual fee next are more likely to cancel than someone scheduled to pay a monthly fee).

Involuntary churn is down. Our big subscriber renewal stretch in September, when ~15,000 annual subscribers came up for renewal, went about as smoothly as we could have hoped, in part because involuntary churn (i.e., credit card failures in Stripe due to expired cards, bank declines, etc.) was down several hundred basis points versus last year. We worked with Lede to optimize some back-end stuff related to storing and updating credit cards over the summer, but to be honest the churn reduction was better than we were expecting so we’re wondering if we’re benefiting from some exogenous effects. If you’re seeing something similar in your subscription business, or you work in payments, we’d be glad to compare notes and try to figure out what’s actually moving the needle.

Normal Gossip subscribers are moving further into the Defector fold. One of our goals for the year was to get Normal Gossip subscribers to read more blogs. We’ve made two major interventions to these ends: First, the NG team launched a biweekly newsletter back in April. (Register here!) The newsletter reaches 10k+ email addresses, roughly a 50/25/25 blend of non-subscribers, NG-specific subscribers, and Defector core subscribers. After a couple of months, we ran the numbers and gave guidance that things were going extremely well, with 70-80% open rates and 10-15% click-through rates, but that we should expect some regression back toward less deranged levels of engagement. Fools! The most recent edition of the newsletter reached a couple thousand more readers than the ones earlier this year, and its open rate remains parked in the mid-70s.

Second, we introduced onboarding appeal emails from Kelsey and Alex to the Normal Gossip product. Doing so appears to have roughly doubled first-month engagement with Defector blogs among Normal Gossip subscribers. (It is, of course, relatively easy to double a small number, but we are nonetheless encouraged by the progress.)

Our homepage isn’t back, because it never left. We don’t really have an opinion on if homepages are back in general, because in our little corner of publishing, Defector’s subscribers have always engaged with the homepage. Nearly 75% of subscriber visits to Defector begin with an organic homepage view. We classify a further 15% of subscriber sessions as organic, where a subscriber clicked through on a newsletter link, loaded a favorite category or tag page, or picked back up where they left off on a blog. Thankfully, our existing subscribers are largely not using search or social media to find their way to Defector blogs.

One year of crosswords: Crosswords, launched in October 2023, have been a nice additional benefit that a good number of subscribers are taking advantage of. Roughly 10% of our subscribers are regular solvers, and crosswords comprise around half the Defector diet of several hundred subscribers.

The (figurative and literal) taxes of operating across states: A hidden cost of being a truly remote company is the administrative burdens for staying in compliance with various state authorities in every state in which you have at least one employee, particularly around payroll taxes and paying into unemployment insurance. The PEOs are generally good at staying on top of these issues, but the moments of handing-off between PEOs have proven to be difficult periods for us, as balls get dropped and log-ins get lost. It took over 90 days to gain access to Defector’s unemployment account with the state of Ohio (special shout out to Carolyn at the Department of Job and Family Services, for finally figuring out that all of our past access recovery attempts had been futile because the account had been created without any email address attached). And we’re still waiting for a five-figure refund of overpaid payroll taxes in a different state, stemming from a botched transition of third-party administrator access in 2022.

How to help other emerging independent media collectives: One of the great joys of working at Defector is the opportunity to help groups of journalists who are hoping to get their own ventures off the ground. Most of the conversations don’t go anywhere, but it’s incredibly gratifying whenever a new employee-owned publication does launch. We’re always happy to offer what support we can, but our responsibilities at Defector mean we can’t really lend more than an hour or two at a time, and it’s frustrating that we’re not able to provide more systematic, rigorous guidance. We’ve had one-off conversations with foundations about what it might look like to build out programming and shared services infrastructure for emerging independent publications, but we’ve failed to find any traction. If you’re a wealthy person looking to put a chunk of money into the space, please do reach out—we have all sorts of interesting ideas for how to support new publications. You can do a lot of good with a relatively small amount of money, and maybe we can, like, name a program or fellowship after you.

Wrapping up

Wow, we really thought this Annual Report was going to be a lot shorter than past editions, but it turns out we still had plenty to say this year. There were various places in the report where we asked for input; please do take those requests seriously and reach out to jasper@defector.com or sean@defector.com if you might have useful thoughts for us.

Thank you for your continued support of Defector. See you tomorrow.