40 hours in the cockpit, never less. 20 hours must be with an instructor. That is the minimum bar to even be tested to earn your private pilot license in the US. You cannot watch a video about landing and hop in a plane and see how you do.
The process is a sequence of increasingly difficult gates: ground school, dual instruction with a certified flight instructor, solo circuits, a solo cross-country, a written knowledge exam, and then a check ride with a designated examiner who watches you fly in real time. Every gate is active, follows a track, and is supervised. The format of flight training is not a quirk of aviation culture. It is the reason trained pilots can actually fly safely.
Without the progressive challenge structure and without an instructor and a community, most student pilots would stop somewhere around the first solo attempt and never return to the airport.
Unfortunately, many online course creators are building the opposite of that: a library. A self-paced collection of recorded videos, available whenever, completable whenever, which for most buyers means never. AI is making content itself increasingly a commodity. Sure, learning requires content, but great learning also requires stewardship, a community, and a sequence of increasingly difficult challenges that leads you to competence.
This Week’s Finds
-
Story 1
Self-paced completion rates fell to 3-5 percent — and paid challenges are completing at 70-80 percent
CommuniPass, Course Creator Pivot 2026 ↗
CommuniPass, Course Completion Rate Problem ↗
Research tracking course platforms in 2026 shows self-paced online courses have dropped below 5 percent completion. The same data shows paid challenges — 14-to-30-day cohort programs with daily prompts, priced between $49 and $199 — completing at 70 to 80 percent. One documented case study followed a marketing course creator whose annual revenue slid from $180,000 to roughly $60,000 over three years as static course sales declined, then recovered to $837,000 after restructuring her primary offer as paid challenges. The 14x completion gap is not a content quality gap. It is the difference between a product people finish and a product they buy and park.
-
Story 2
Klaviyo expanded its Anthropic integration so AI agents can now run email campaigns autonomously
Klaviyo, Agentic Marketing Workflows ↗
Business Wire, Klaviyo Expands Integration with Anthropic ↗
On May 7, Klaviyo announced an expanded integration with Anthropic that connects Klaviyo’s customer data to Claude through a Model Context Protocol server. In practical terms, Claude can now pull campaign performance data, audit underperforming flows, generate email copy from segment data, and save finished assets to client folders — chaining those tasks without a human queuing each step. For agencies managing multiple Klaviyo accounts, what was a half-day reporting task becomes a scheduled overnight job. Klaviyo describes the approach as model-agnostic, and the new Query Metric Aggregates tool exposes raw event data rather than pre-summarized dashboards, which makes AI-driven analysis substantively useful rather than generic.
-
Story 3
LinkedIn's new 360Brew algorithm is penalizing template and AI-generated content and rewarding narrow-topic consistency
Agorapulse, LinkedIn Algorithm 2026 ↗
LinkedIn replaced its content ranking infrastructure with a system called 360Brew and the reach numbers are showing up clearly. Organic reach dropped roughly 50 percent year-over-year for accounts posting generic or AI-generated content. Accounts that post consistently on two or three specific topics over a 60-to-90-day window are being indexed by 360Brew as credible voices in those topics and getting stronger distribution than before the change. Engagement pods — a long-standing workaround where groups of accounts boost each other’s posts — are being actively detected and neutralized. For coaches and consultants using LinkedIn to generate leads, the algorithm change has one clear implication: specificity and consistency on a narrow topic set outperform volume and variety.
3%
Average completion rate for self-paced online courses in 2026. Paid challenges, cohort programs with daily accountability and a fixed endpoint, average 70–80% completion — a 14x gap that runs through every downstream metric.
Three stories this week, across different platforms and different mechanisms, but the same idea underneath each one: the structure of how you deliver something determines whether people finish it, whether it works for them, and whether they come back.
The course completion data is the loudest signal. At 3 to 5 percent completion, a self-paced course is not really a product people consume. It is a product they buy and park. The purchase feels like progress. The actual transformation the course promises does not happen. Which means the testimonial, the referral, the upsell to your next program — none of that follows either. The revenue math gets punished fast when you trace it: if only 5 out of 100 buyers finish, you have 5 people who can speak to results and who are ready for your next offer. A paid challenge at 75 percent completion gives you 75 people in that position.
The aviation training parallel is not a stretch. Flight training is structured to produce completion because it has to. Progressive challenge gates, dual instruction with live feedback, solo milestones with real stakes, a check ride where you either perform or reschedule. Every stage requires active participation. Nothing is passive consumption. That architecture is why trained pilots can fly airplanes safely. The same architecture, applied to a 14-to-30-day cohort program, produces the same results in course completion.
Here is what to do with this information this week. If you are running a self-paced course as your primary offer, do this calculation: how many buyers from the past 12 months can you name as documented success stories? If that number is under 20 percent of total purchases, your format is quietly bleeding your business. The fix is not a better module, a new platform, or a stronger sales page. The fix is adding a cohort layer to what you already have. A 14-day paid challenge on your core topic, priced between $97 and $197, with a live kickoff call, daily prompts delivered by email or in a community channel, and a midpoint check-in call. That structure alone, without changing a single piece of content, will move your completion rate from single digits into the 60 to 70 percent range.
On LinkedIn: if you have been posting without a deliberate topic system, pick two specific topics tied to your expertise and post four times a week for 60 days. That is the threshold at which 360Brew begins recognizing you as a credible voice. Track engagement at day 30 and day 60.
On the Klaviyo and Claude integration: if you use Klaviyo for your email list, the new MCP connector is worth looking at specifically for flow audits. Ask Claude to pull your flow performance data and flag sequences where open rates drop more than 20 percent between the first and third email. That drop-off is your leaky bucket. Fix those two or three flows before building anything new. Klaviyo’s own data shows automated emails drive 30 percent of all email-generated sales despite being only 2 percent of total sends. Getting your existing flows right is higher leverage than adding more sequences.
IMG’s Take
The conversation in the course creator and coaching space right now is almost entirely about AI tools for content production and almost entirely missing the structural problem underneath. AI can write a better module script, a better sales page, a better email sequence. It cannot make someone sit down and watch. What produces completion is accountability baked into the format, a deadline that matters, a community watching progress, and a milestone that has weight. That is challenge architecture. The tools that support it will keep changing. The architecture does not.
IMG Courses has watched this play out directly. The programs generating the most consistent student outcomes and the most referrals are not the ones with the highest production budgets or the most sophisticated AI-generated content. They are the ones where students have a clear endpoint, a reason to show up today, and something that marks completion as distinct from mere enrollment.
The Klaviyo and Claude story deserves attention, but not as hype. AI marketing automation produces real value when it runs against specific data and generates outputs a human can review and ship. It fails when it generates summaries nobody reads. The distinction is the same as it always has been: specific beats generic. Claude pulling your actual flow metrics and identifying where your welcome sequence loses 40 percent of opens is useful. Claude writing a list of email marketing tips is not. Point it at the data you already have. Let it do the analysis work. You do the judgment calls.
Pick the format that produces completion. Build automation around the data you already own. Both of those moves will outperform any content volume strategy running in 2026.
Have you made the pivot from self-paced to cohort or challenge format? What did your completion rates and revenue per student look like before and after? Bring your ground-level data to the vault forum — real numbers from real creators are what move the industry forward.
Join the IMG Community →
— Tim Nichols & The IMG Team
Sources cited in this edition
- CommuniPass, Course Creator Pivot 2026 (completion rate data, $60K to $837K case study): CommuniPass
- CommuniPass, Course Completion Rate Problem (3-5% self-paced, 70-80% challenges): CommuniPass
- Klaviyo, Agentic Marketing Workflows with Klaviyo and Anthropic Claude (May 7 MCP integration): Klaviyo
- Business Wire, Klaviyo Expands Integration with Anthropic (announcement): Business Wire
- Agorapulse, LinkedIn Algorithm 2026 — 360Brew rollout and reach data: Agorapulse
