The problem your training dashboard doesn't show
Every year, French companies invest heavily in training. Per DARES (2024), companies with 10+ employees reported around €10 billion in direct training spend in 2023, between 1% and 2% of payroll on average. For an organization with 300 employees and a €15 million payroll: between €150,000 and €300,000 invested every year.
These budgets fund authoring tools, LMS licenses, design time, providers, trainers. Dashboards display reassuring metrics: high completion rates, satisfying scores, rising training hours. On paper, it works.
But here's what those dashboards don't show: what your learners have really retained ten days after clicking Complete. The scientific answer has been known for 140 years. It's robust, replicated, universal. And it's systematically ignored when designing training programs.
The forgetting curve: 140 years of results training ignores
In 1885, German psychologist Hermann Ebbinghaus published Uber das Gedachtnis (On Memory). He documented a fundamental finding: retention of information decays exponentially over time with no reactivation.
By Day 1: about 70% forgotten. By Day 7: 90%. By Day 30: less than 5%.
This finding, the forgetting curve, has been replicated hundreds of times in very different contexts. The exact numbers vary, but the shape of the curve is universal and undisputed. It's not a flaw; it's memory's cleanup mechanism, clearing out information that hasn't been reactivated.
Cards addresses this exact mechanism with its built-in Learning Routine® anchoring engine.
Why workplace training makes forgetting worse
In a workplace setting, several factors make the Ebbinghaus curve worse.
No recall context: in academic learning, regular exams force reactivation. At work, training ends once the module is complete.
Proactive interference: the working learner is exposed to a constant stream of new information that speeds up forgetting.
Delay before application: several weeks can pass between the end of training and the first chance to apply it.
What the research says
Learning psychology has produced robust findings on the conditions that support long-term retention. These findings are published in major journals but remain largely ignored in workplace training practice.
Spacing Effect
Several short, spaced sessions produce markedly better retention. Cepeda et al. (2006) meta-analysis of 254 studies and 14,000 participants: +200% to +400% retention.
Testing Effect (active recall)
Roediger & Karpicke (2006): +50% retention at one week for learners who were tested vs. those who reread. Quizzes anchor, not just assess.
Interleaving
Alternating different types of content during reminders improves retention. Rohrer et al. (2014): +43% vs. blocked practice in mathematics.
Generation Effect
Actively producing an answer in your own words is more effective than passive reception. Slamecka & Graf (1978) documented this, and hundreds of studies have confirmed it since.
The converging conclusion: the most effective training regularly forces the learner to actively retrieve information from memory, at spaced intervals, in varied formats. It's not longer or visually fancier. It's more regular, shorter, more active.
This logic of regular activation is also what sets Cards apart from a classic LMS: see the Cards vs. LMS comparison.
Calculating the cost of forgetting
Calculating the cost of forgetting requires linking three line items, the third of which is structurally absent from almost every training budget.
Content production cost
Instructional design, authoring tool, revisions, business validation. The historical industry benchmark, Bryan Chapman (Brandon Hall Group, 2023), estimated an average of 197 hours (28 days) of work to produce one hour of standard e-learning. That ratio is now largely outdated: per the 2026 ISTF survey (460 training professionals), 33% of designers now use AI in their instructional practice, and AI has become the #1 priority for 30% of training departments. In practice in 2026, observed ratios range between 60 and 100 hours per hour of e-learning, or about 14 days of work instead of 28. With Cards and its 4 AI agents, design time is distributed directly to subject-matter experts: 5 microlearning courses produced in 10 to 45 minutes.
A considerable time saving, with no compromise on quality.
Delivery and management cost
LMS license, technical maintenance, module updates, reporting and administrative management. Between €3 and €15 per active learner per year depending on the LMS.
Cost of forgetting ← the invisible line item
This line item appears in no training budget. If 70% of content is forgotten by Day 1, actual memorized value represents less than 30% of the investment. For a €5,000 module: production cost €5,000, memorized value €1,500.
Cost of forgetting: €3,500 evaporated, 70% of your investment.
An organization investing €200,000/year in training with no anchoring produces real memory impact of €40,000 to €60,000 at 30 days. The remaining €140,000 to €160,000 is the cost of forgetting, invisible in every dashboard.
"Forgetting" is the biggest line item in your training budget. It's just invisible in every dashboard.— Matthieu THOMAS, co-fondateur Cards · Dir. Learning & Devlopment
Three concrete scenarios
Scenario 1: GDPR compliance training (200 employees)
Annual GDPR training · 200 employees
Scenario 2: Sales onboarding (20 new hires/year)
Sales onboarding program · 20 new hires / year
Cards specifically addresses onboarding programs.
From the forgetting curve to the retention curve
The answer to the forgetting problem is in Ebbinghaus's own work: by reactivating information before it's forgotten, you reset its forgetting clock to zero. This is the principle of spaced repetition: Day 1, Day 3, Day 7, Day 14, Day 30.
Formation initiale
The learner completes the Cards microlearning course. Immediate retention close to 100%. The memory trace is fresh.
First automatic reminder (Learning Routine®)
Without anchoring: 70% already forgotten. With Cards Learning Routine®: a 2-3 minute reminder reactivates the memory trace before mass forgetting sets in.
Adaptive reminders on weak spots
Without anchoring: 90% forgotten. With Learning Routine®: 2-3 spaced reminders, focused on each learner's weakest points.
Anchoring consolidated: retention measurement
Without anchoring: less than 5% retained. Logged as "Complete" in the LMS, but the content is gone. With Cards Learning Routine®: 80%+ retention. The knowledge is anchored, ready to use on the job.
What anchoring actually changes
5 priority actions to reduce the cost of forgetting
1. Identify your training with critical retention stakes
Not all training carries the same stakes. Prioritize the courses where forgetting has direct operational consequences: regulatory compliance, safety procedures, product training for sales teams, onboarding for complex processes.
2. Break your existing modules into microlearning units
A 60-minute module is structurally less well anchored than six 10-minute modules spread out over time, even with identical content. With Cards AI, this work can be done quickly from an existing PDF or PowerPoint.
3. Systematize post-training reminders
Automated through Cards and its Learning Routine®, the reminder process scales to hundreds of learners with no extra effort. The Day 1 testing effect is especially powerful: it kicks in exactly when the curve drops the fastest.
You can also run this process from your enterprise AI via the Cards MCP connection.
4. Change your management indicators
Adding Day 30 retention rate to your dashboard repositions your training function around real impact rather than consumption. Cards is one of the only tools on the French-speaking market to natively expose this metric.
5. Calculate and present the cost of forgetting to your leadership
Translating the phenomenon of forgetting into financial terms is the most effective way to win leadership's support. A figure of €60,000 to €120,000 in lost training value per year makes investing in an anchoring tool immediately justifiable.
Sources and references
- Ebbinghaus, H. (1885). Uber das Gedachtnis. Duncker & Humblot, Leipzig.
- Cepeda, N.J. et al. (2006). Distributed practice in verbal recall tasks. Psychological Bulletin, 132(3), 354-380.
- Cepeda, N.J. et al. (2008). Spacing effects in learning. Psychological Science, 19(11).
- Roediger, H.L. & Karpicke, J.D. (2006). Test-enhanced learning. Psychological Science, 17(3).
- Rohrer, D. et al. (2014). The benefit of interleaved practice. Psychonomic Bulletin & Review, 21(5).
- Chapman, B. (2023). How Long Does It Take to Create Learning? Brandon Hall Group Research.
- DARES (2024). La formation professionnelle continue en entreprise — résultats 2023. Ministère du Travail.
- Kapp, K.M. & Defelice, R.A. (2019). Microlearning: Short and Sweet. ATD Press.