In 2007, Microsoft made a discovery that should have changed every software company's strategy: approximately 90% of the feature requests they received for Office were for features that already existed in the product.
Users weren't asking for innovation. They were asking for education.
Microsoft's solution was the Ribbon UI — a visual redesign to make existing features more discoverable. But the underlying problem wasn't unique to Microsoft. It's a structural problem that gets exponentially worse as SaaS companies add more products to their platform.
Today, public cloud companies spend approximately $29.5 billion on features that may rarely or never be used (Pendo). Not because the features are bad. Because nobody taught customers they existed.
The Median Feature Adoption Rate Is 6.4%
Pendo's 2024 Software Benchmarks analyzed feature adoption across their customer base and found the median feature adoption rate is 6.4%. For every 100 features a company ships, only 6.4 drive 80% of click volume.
The top 10% of products achieve 15.6% — still meaning 84% of features get minimal use.
This isn't a product quality problem. Userpilot found that 68% of feature adoption failures stem from discoverability rather than inherent value gaps. Users can't find the features, not that the features aren't good.
The 80% waste stat from Pendo's 2019 Feature Adoption Report is even more stark: 80% of features in software products are rarely or never used. Pendo analyzed 615 subscriptions and categorized usage into frequent, moderate, rare, and never.
Meanwhile, organizations waste $21 million annually in unused SaaS licenses. For large enterprises, that number reaches $127 million (Zylo 2025 SaaS Management Index). And the waste increased 14.2% year-over-year.
Why Multi-Product Suites Make It Exponentially Worse
Here's where most people misunderstand the problem: they think adding products creates linear education needs. It doesn't. It creates exponential ones.
If you have one product, you need education for that product.
Add Product B, and you need education for Product B plus 'how A and B work together.'
Add Product C, and you need education for Product C plus A+C, B+C, and A+B+C.
The formula is roughly 2^n - n - 1 additional cross-product education paths for n products. For HubSpot's 5 major Hubs (Marketing, Sales, Service, Content, Commerce), that's 26 possible cross-product education combinations beyond individual product training.
This is the product proliferation tax. Every product you add doesn't just create its own education burden — it multiplies the education needs across every existing product.
HubSpot shipped 400+ product updates in 2024 alone. That's roughly 33 new things per month that customers theoretically need to learn about, evaluate, and potentially adopt. Their R&D spending reached $779 million in 2024 — a 2,300% increase over ten years.
The shipping velocity is accelerating. The education infrastructure is not.
The Numbers That Prove the Gap Is Growing
Six new SaaS applications enter the average organization each month (Zylo 2025). Enterprise environments now manage 275 SaaS apps on average, with IT overseeing just 26% of spend — a 6.4% drop over 2024.
The average employee switches between 13 different SaaS applications throughout their work day (Asana Work Innovation Lab).
Organizations use only 47% of their SaaS licenses. 53% go unused in a given month, up 7% year-over-year.
Meanwhile, 23% of companies report customers use half or less of product functionality (Intellum/Forrester). Over 28% of customers aren't even offered training. And only 29% of a customer's user base engages in training annually (SkillJar 2022).
So: 80% of features unused, 53% of licenses wasted, 71% of customers never trained. And companies are shipping faster than ever.
The Three Giants Show the Scale
HubSpot: 39% of Pro+ customers subscribe to four or more hubs. 61% of new Pro+ customers land with multiple hubs from day one. HubSpot Academy has trained 450,000+ professionals. Yet 89% of unified platform customers who use it daily generate 107% more leads, 35% more deals, and 28% more ticket resolution — proving that cross-product education works, but also that only the deeply trained users get there.
Salesforce: 20+ clouds, 48+ certifications, 7,000+ AppExchange apps. 94% of salary survey respondents use Trailhead. The ecosystem is projected to generate $1.6 trillion in new business revenues by 2026. But achieving Ranger status requires 100 badges and 50,000 points — that's the education investment needed to truly understand the platform.
Adobe: Committed $100M+ in 2024 for education initiatives. Targets 30 million learners by 2030. Offers certifications at three levels across three job roles, multiplied across Creative Cloud, Experience Cloud, and Document Cloud. The education surface area is staggering.
These companies can afford to invest hundreds of millions in education. What about the 96% of B2B SaaS companies that can't?
The Cognitive Science Explains Why It Keeps Getting Worse
This isn't just a business problem. It's a cognitive architecture problem.
Working memory holds approximately 4 ± 1 items at a time (Cognitive Load Theory, Sweller). Every additional feature, menu item, or product in a suite competes for this extremely limited capacity.
Hick's Law: the time to make a decision increases logarithmically with the number of choices. When users face too many options, they experience decision paralysis and abandon the task entirely.
Columbia Business School found that after the optimal point (usually 3-4 key benefits), additional information actually decreases conversion rates. More features presented at once = fewer features adopted.
Decision fatigue compounds throughout the day. A study of traffic court judges showed the longer the time since a judge's last break, the lower the probability of case dismissal. The quality of decisions deteriorates over time.
Now apply this to a SaaS user facing 275 apps, switching between 13 of them daily, with 80% of features they've never been trained on. Their brain isn't lazy. It's operating at maximum capacity with no education infrastructure to offload the cognitive burden.
From the nervous system's perspective (Savic 2018), every unfamiliar interface element registers as a micro-threat. The amygdala flags uncertainty. Cortisol rises. Working memory narrows. The user retreats to whatever feels safe — the 6.4% of features they already know.
The Paradox: Multi-Product Users Are 66% More Valuable — And the Least Served
Here's the paradox: customers with multiple product relationships have a 66% higher retention rate (Gartner). Later-stage SaaS companies with the lowest churn were those that cross-sold to about one-third of their customers (McKinsey/Gainsight SaaSRadar).
Multi-product users generate 39.2% of new revenue through expansion (SaaS Capital).
These are the most valuable customers. They retain better, expand more, and generate significantly more revenue.
But they also need the most education. HubSpot's 39% of Pro+ customers on 4+ hubs need to understand not just each hub but how they interconnect. Their average subscription revenue is $11,600 — these are high-value accounts that need deeper education than any one product can provide.
And here's where the paradox bites: 52% of companies lack tools to build training resources (Intellum/Forrester). 42% lack personnel. Only 4% describe their program as 'formalized, scalable, and curriculum-based.'
The most valuable customer segment needs the most education. Almost no one is delivering it.
Education Closes the Gap — When It Exists
Customer education programs increase product adoption by 38.3% (Forrester/Intellum 2024). They deliver 372% ROI with a 7-month payback. 96% see positive returns. 35% increase in lifetime value per trainee.
The data is clear. Education works.
HubSpot's Content Hub attach rate to Marketing Hub went from 13% to 54% in a single year — demonstrating that intentional cross-product education can dramatically close the adoption gap.
Users who engage with new features within their first week show 3.7x higher 6-month retention compared to users who delay feature discovery beyond 30 days (SaaS Factor).
After Pendo helped improve feature discovery, customers saw a 50% increase in daily feature use and unused features declined by nearly 25%.
Businesses plan to nearly triple their spending on customer education between 2024 and 2026 (Intellum). Gainsight acquired Northpass (2023), Staircase AI (2024), and Skilljar (April 2025) — consolidating education into the customer success platform. The market is recognizing that education isn't optional.
The Cost of the Gap
$29.5 billion on features that may never be used. $21 million per year per organization in wasted SaaS licenses. 80% of features ignored. 53% of licenses unused. 68% of adoption failures from discoverability, not value.
And this: 90% of feature requests were for features that already existed.
Microsoft learned this in 2007 and redesigned their UI. Almost two decades later, the rest of the industry is still shipping features without teaching them.
The product proliferation tax compounds with every release. Every new feature shipped without education widens the gap. Every new product added without cross-product training multiplies the confusion.
The companies that figure out customer education will keep their most valuable multi-product users. The ones that don't will watch 93.6% of their features gather dust while their support teams field requests for things that already exist.
That's the multi-product education paradox. And it's getting worse every quarter.
---
Sources:
Microsoft / Jensen Harris blog (2007, via John D. Cook); Pendo 2019 Feature Adoption Report (n=615); Pendo 2024 Software Benchmarks; Zylo 2025 SaaS Management Index; Userpilot; HubSpot INBOUND 2024 / Q4 2024 Earnings / Spring 2025 Spotlight / GetLatka; Salesforce Ben; Adobe Press Release October 2024 / Adobe Blog; Intellum / Forrester 2024 (n=300); SaaS Academy Advisors; SkillJar 2022; SaaS Capital; Gainsight CS Index 2025; Asana Work Innovation Lab; BetterCloud 2025; Gartner; McKinsey/Gainsight SaaSRadar; Columbia Business School; Cognitive Load Theory (Sweller); Hick-Hyman Law (1950s); Decision Fatigue (Baumeister); Savic 2018.