Business Training

How do you design a new hire
onboarding program?

Last updated: 2026-07-14

Onboarding is a loop, not a checklist. And once it ends, the variable is not total volume — it is refresh frequency.

How often do you actually need to publish?

Volume is not the variable. Refresh frequency is. A company holding 1,000 courses still fell below 10% MAU when new uploads stopped for two months, while companies publishing even one item a week held 15–25% off-peak (/en/data-report). Across 107 companies, not one sustained 70% MAU while publishing less than weekly. A mid-size pharmaceutical company with about 3,000 employees built a standing habit out of lunchtime microlearning (/en/content-efficiency).

Onboarding time fell about 70% and offline training costs about 63% — but both figures come from fewer than five companies, so treat them as indicative rather than settled. Level 4 business impact was demonstrated at fewer than five of 107 companies, which is worth saying plainly: adopting an LMS does not generate that evidence on its own. For the vendor-neutral view, see our 8 criteria for choosing an enterprise LMS.

New Hire Onboarding
~70% Faster

Mission-based self-paced learning gets new hires field-ready sooner. Sample of fewer than five companies — treat as indicative, not settled.

previous onboarding30day

TC After9day

70% reduced ✓

Publish Weekly,
or Engagement Dies

Even 1,000 courses could not hold engagement once uploads stopped for two months. Companies posting at least weekly held 15–25% off-peak.

<10% ↓ after 2 months of silence

Weekly uploads: 15–25% off-peak MAU

Video First + Field Practice:
Flipped Learning in Action

Global Tire Co. has staff learn via video first, then practice in the field — flipped learning that boosted training impact.

flipped learning process

STEP 1

video learning

→

STEP 2

field practice

→

STEP 3

quiz verification

39,000 Users
on a Dedicated Mobile App

A major delivery company trains 39,000 people nationwide in the field through a single dedicated app.

39K

User

Dedicated app

Nationwide

field coverage

Can onboarding actually
be made shorter?

New hire onboarding

Requiring course completion alone just moves classroom training onto a phone. What worked was a loop: assign a mission, get a peer response, build belonging. Profile setup, a self-introduction post, and assigned courses were bundled into missions. Across the ~25 companies where this pattern was observed, median content completion for new hires was 85%+ (range 70–95%, n≈25). Onboarding is typically designed to run 2–4 weeks. Source: /en/data-report

New-hire completion 85%+ (n≈25)

OnboardingConcept illustration

Install app
Today's mission
Expert know-how
Completed

Sales capability building

A major Korean insurer recorded 5,236 learning posts from a single Live Sessions lecture series. Sales reps use lunchtime microlearning, maintaining MAU at 85%+ during mandatory training periods. Top performers' know-how spreads in real time.

MAU 85%+ / 5,236 Live Session posts

SalesConcept illustration

Sales

On the commute

On a break

At work

Field worker training

No PC needed — a smartphone is enough. A major delivery company trains 39,000 users via a dedicated app. Each content piece takes under 5 minutes, so staff learn during short breaks between tasks.

Under 5 min per item / 39,000 users

In the fieldConcept illustration

In the field

Safety training

Work manual

Short-form

How do you know content
is actually being consumed?

Onboarding period reduced

Across the ~25 companies running mission-based onboarding, median content completion for new hires was 85%+ (range 70–95%, n≈25). Onboarding is typically designed to run 2–4 weeks. Among companies that ran a cost comparison, offline training costs fell about 63% — but that sample is fewer than 5 companies, so treat it as indicative rather than settled. Source: /en/data-report

New-hire completion 85%+ (n≈25)

CorrelationObserved data

<20%Under 3 / month

50%+10+ / month

n=75 · Correlation does not establish causation

Field job skills microlearning

35+ companies run role-based training programs for field teams, reaching 85%+ MAU during mandatory training periods. Lunchtime microlearning for sales reps has reached 95% MAU. Publishing at least one session a week keeps MAU in the 70-95% range.

MAU 70-95%

Course volumeConcept illustration

Required training

Job skills

Keep learning

Cost savings impact

Instructor fees, venue rental, and travel costs disappear. Among companies that ran a cost comparison, offline training costs fell about 63% — but that sample is fewer than 5 companies, and a company already largely online will see far less. One large insurer that moved its printed materials to E-Book cut its material production budget by 40%. Calculating from your own offline spend is more accurate.

~63% lower than offline (n<5)

Cost structureConcept illustration

In-person training

Content reuse

What if you have no dedicated content team?

To maintain a publishing cadence, each item must take less time to produce.
Evaluate an AI LMS by the time it actually saves your team.

STEP 1

Add source materials

Start with internal documents, PDFs, or web URLs. Existing PPT and PDF training materials become your source.

STEP 2

Generate a curriculum

Build an outline around learning goals, then review the sequence and scope.

STEP 3

Create learning pages

Generate 500–4,000 characters per page, check spelling, choose from five tones, add AI images, and translate into 14 languages.

AI content creation

Documents into courses

Create curricula and learning pages from URLs and files, with translation into 14 languages and AI image generation.

AI ShortClass

Long videos into short clips

Automatically extract short learning clips from existing long videos to reuse content more often.

ReFlash

Scripts into videos

Create videos with images, captions, and narration from scripts or PDFs.

AI curation

Relevant to each learner

Recommend content based on behavior, roles, and interests so learners see material relevant to their work.

AI learning assistant

Answers within the course

A closed AI environment answers course questions using internal training materials. Customer knowledge assets are not used to train AI models.

AI Admin

Ask about data in natural language

Ask about learning data, receive reports or CSV exports, and follow up with reminders.

Lessons from AI adoption. One company found errors in details such as prices and conditions after adding all its content at once. It switched to a category-by-category rollout validated through a one-month pilot. Limit the reference scope and validate answers. For an example of automated field inquiries, see mobile learning for frontline teams.

How many courses do you need, and how often must you publish?

Drawn from 107 companies over 35 months. Where the sample is thin, we say so in the row.

Table A. Content volume, frequency, and engagement

A baseline you can hold a vendor's proposal against. The figures that held and the ones that collapsed are both here.

Metric	Value	Sample & conditions	Source
Publishing frequency vs. engagement	10+/month → MAU 50%+	100+ companies · fewer than 3/month → MAU under 20%	Operating data report
Falsifiable finding	0 companies	Of 107, none held 70% MAU while publishing less than weekly	Operating data report
Course volume vs. off-peak MAU	1–5% / 10–20% / 30–50%	Compliance-only / 500+ courses / 2,000+ courses	Operating data report
Volume without refresh	under 10% MAU	Even with 1,000+ courses, when uploads stopped for 2 months. Weekly uploads held 15–25% off-peak	Operating data report
Onboarding duration	~70% shorter	Sample of fewer than 5 companies — indicative, not settled	Operating data report
Cost vs. offline training	~63% lower	Sample of fewer than 5 companies. If your training is already largely online, expect far less	Operating data report
Microlearning in practice	~3,000 employees	Mid-size pharmaceutical company · lunchtime microlearning habit	Customer stories
Depth of consumption	658 pages per person	Bluebell Korea · 85,000 cumulative logins, 18,000 learning hours, 1.1M pages	Customer stories
Level 4 business impact proven	fewer than 5 of 107	Adopting an LMS does not generate this evidence by itself	Operating data report

On reducing repeat field inquiries with an AI FAQ, see frontline training. Sources: https://www.touchclass.com/en/casehome · https://www.touchclass.com/en/data-report

Where this sample is limited. These companies chose a mobile-first platform on their own, so the sample carries selection bias. It is not a population statistic. The onboarding and cost-reduction figures rest on fewer than five companies each — they are directional, not benchmarks you should plan a budget around. The relationship between publishing frequency and engagement is a correlation; we did not control for other variables, so it is not proof of cause. And the 500-course threshold is probably less important than what those courses actually are: composition may matter more than count.

Frequently asked questions

The eight questions that come up most often about onboarding and content operations.

How should a new hire onboarding programme be structured?

Fix the sequence so it runs without a manager driving it: a short video, a checklist, and a quiz that confirms the material landed, assigned automatically from the join date. Mission-based self-paced learning lets a new hire progress without waiting for the next scheduled session. Source: https://www.touchclass.com/en/content-efficiency

How much can onboarding time actually be reduced?

Onboarding time has come down by roughly 70% in the cases we have seen, but we will be honest about the evidence: that comes from a sample of fewer than five companies, so treat it as indicative rather than a benchmark to plan against. Companies already running structured online onboarding should expect a much smaller gain than a company moving off entirely in-person training. Source: https://www.touchclass.com/en/data-report

How many training courses do we actually need?

Off-peak MAU tracked course volume: 1–5% for compliance-only libraries, 10–20% above 500 courses, and 30–50% above 2,000. But the number is the weaker half of the story — composition likely matters more than count, and a large library that is never refreshed still collapses. Source: https://www.touchclass.com/en/data-report

How often do we need to publish new content?

At least weekly, ideally 10 or more items a month. Companies publishing 10+ a month held MAU above 50%; those under 3 a month sat below 20%. The hardest number for us to explain away: across 107 companies, not one sustained 70% MAU while publishing less than weekly. Source: https://www.touchclass.com/en/data-report

Can we reduce the cost of outsourcing content production?

Producing content in-house with AI tools removes most per-item agency cost, and the practical shift is that a subject expert can publish directly instead of briefing an external producer. Reported savings against offline training run about 63%, but that comes from fewer than five companies, and a company already largely online will see far less. Source: https://www.touchclass.com/en/content-editor

Does microlearning actually work?

Microlearning works when it fits an existing gap in the day rather than asking for a new one. A mid-size pharmaceutical company with about 3,000 employees built a standing habit around lunchtime microlearning. The mechanism is scheduling, not format: short content still goes unopened if it arrives when people cannot look at a phone. Source: https://www.touchclass.com/en/casehome

How should sales training be run?

Publish to the rhythm of the role. Salespeople are in meetings and in transit for most of the day, so content has to be consumable in three to five minutes, ideally with an audio format for travel. Product updates should land before the customer conversation, not in a monthly training block after it. Source: https://www.touchclass.com/en/content-efficiency

Can we reuse our existing PPT and PDF training material?

Yes. Existing PPT and PDF material can be uploaded and converted for mobile viewing, and can be turned into video or e-book formats rather than rebuilt from scratch. Depth of consumption is measurable — at Bluebell Korea, 1.1 million pages were read, averaging 658 pages per person. Source: https://www.touchclass.com/en/content-editor