Customer education programs for SaaS

Customer education gets funded as a support cost and evaluated as a support cost, which is why it’s usually underbuilt and always under-measured. The better frame: it’s the only content you own that both keeps existing customers and attracts new ones, because “how do I do X in Y” is a query that prospects and customers type identically.

What follows is the decision about which level to build, how to structure the curriculum, whether certification is worth it, and the four numbers that tell you if any of it worked.

Three levels, and which one you should build

They’re not stages of the same thing. They’re different products with different costs and different payoffs.

Level	What it is	Rough build cost	When it makes sense
Help center	Searchable task answers, 100 to 600 articles	1 technical writer, ongoing	Always. Build this first and keep it current.
Academy	Sequenced, role based learning paths with video	1 to 2 FTE plus production, 120k to 300k year one	500 plus accounts, multiple personas, real configuration depth
Certification	Assessed credential with a badge and a pass mark	Adds 80k to 200k plus exam maintenance	Only when employers list the credential in job ads

The help center is unglamorous and it’s where nearly all the value sits for companies under a few hundred accounts. Intercom and Zendesk both built help center products precisely because this is the highest volume, highest frequency need, and a good one deflects tickets immediately rather than in a quarter.

The academy earns its cost when your product has enough surface that a user needs a sequence rather than an answer. Multiple roles, a configuration layer, integrations that require decisions. Below that threshold you’re producing a course catalogue for a product people learn in twenty minutes.

What the four famous programs were each built to solve

Naming them is useful because their goals differed, and copying the structure without the goal is how these projects go wrong.

HubSpot Academy was built as a demand generation asset as much as a customer one. Its inbound marketing certification teaches a methodology that makes HubSpot the obvious tool, and the credential appears in marketing job listings, which gives non-customers a reason to complete it.

Salesforce Trailhead solved an ecosystem problem. Salesforce needed a labour supply of administrators and developers, and Trailhead manufactures that supply while gamifying the path. The credential has genuine labour market value, which is the entire reason it works.

Asana Academy solves adoption inside accounts. Its job is getting a newly bought team to actually use the product in the first 30 days, and the content is organised around that.

Webflow University solved a craft problem. Webflow needed designers to learn concepts they’d never had to think about, so the content teaches web fundamentals as much as the tool, which is why it pulls in people who aren’t customers yet.

Four programs, four different jobs. Decide yours before you pick an LMS.

Curriculum design: by role, not by feature

The default mistake is organising education around your product’s navigation. Users don’t have feature shaped problems. They have job shaped problems, and the feature they need is an implementation detail they don’t yet know the name of.

Structure paths by role and by stage within that role. A typical shape for a product with three personas:

Administrator, first week: set up the workspace, configure permissions, connect the two integrations most accounts need. Administrator, first quarter: reporting, governance, user lifecycle.

Everyday user, first day: the one workflow that delivers value, taught end to end. Everyday user, first month: the three next most common workflows.

Executive or sponsor: one 12 minute path covering what to look at and how to tell if the rollout is working. This path is almost always missing and it’s the one that protects the renewal.

Keep units to 5 to 12 minutes with a single stated outcome each. Self paced completion rates for longer courses are poor, and a learner who abandons a 45 minute course at minute 20 has gained nothing, while a learner who does three of eight short lessons has gained three things.

Certification economics, honestly

Here’s my position. Certification only pays when employers value the badge, and almost nobody checks this before building.

The test is simple and takes ten minutes. Search job listings for your product name. If hiring managers are asking for people who know your tool, a certification has labour market value and learners have a career reason to finish it. If they aren’t, your certification is a completion certificate with an exam attached, and it’ll cost you 80k to 200k to produce plus ongoing exam maintenance every time the product changes.

The costs people forget: exam item writing is a specialist skill, questions need refreshing as the product evolves, proctoring for anything high stakes costs money per candidate, and a credential that expires needs a renewal pathway or it quietly dies. Budget maintenance at roughly a third of build cost annually.

What to do instead if the job listing test fails: ship short task based lessons, give a simple completion record, and revisit certification when your product starts appearing in job descriptions. That’s the honest sequence and it saves most companies a wasted year.

Build versus buy, and the SEO decision inside it

LMS options fall into three groups: dedicated customer education platforms, general purpose LMS products repurposed for customers, and building on your existing CMS.

For most SaaS companies under 5,000 accounts, building on your CMS is genuinely competitive. You need video hosting, a progress record, and content pages. A dedicated platform earns its fee when you need cohort management, assessments, certificates and reporting at scale.

The decision that costs the most if you get it wrong: where the content lives. Many platforms default to hosting on their own subdomain, which means every link earned and every ranking achieved accrues to a domain you don’t own. Insist on a subfolder of your main domain, such as yourdomain.com/academy, via custom domain mapping or a reverse proxy. Check this in the sales process, not after migration, because moving later costs you the rankings you built.

The SEO upside is real and is why marketing should care about this at all. How-to queries for your product category are searched by prospects as well as customers, and an academy indexed on your domain compounds with the rest of your content. That is the mechanism described in the retention content system.

Measuring it against things that matter

Completion counts are activity. Four outcome metrics, each compared between trained and untrained accounts.

The organisational question

Customer success owns the outcome. Marketing owns the publishing surface, the SEO and the promotion. One named owner in CS or enablement, with a committed marketing partner, works. Even split ownership between the two functions produces content that ranks for nothing and teaches nobody, which is the most common way these programs die quietly.

Budget realistically. A serious academy is 1 to 2 full time people plus video production, roughly 120k to 300k in year one depending on whether production is in-house. That’s a real number and the program should be defended with the four metrics above rather than with completion counts.

Education also feeds the rest of customer marketing. Trained power users are your best candidates for a customer advocacy program and your most willing participants in a customer reference program, because they’ve already invested effort in your product. The same group populates a user group program. Run the cost of churn calculator to size what a two point renewal improvement is worth before you argue for the headcount, and use the customer story interview questions when you harvest stories from accounts that trained.

What to do next

Run the job listing test this week. Search for your product name in job ads and count the results. If the count is near zero, drop certification from the plan, fix the help center, and build six role based lessons of ten minutes each. Then set up the trained versus untrained comparison before you publish, because retrofitting the measurement six months later is how good programs lose their funding. Wider context sits in SaaS customer marketing, with worked cases in SaaS customer marketing examples and a guided build in the customer marketing sprint.