PTC University — Learning Connector

Read the contract before you touch the screens.

PTC sold its software on perpetual licenses: pay once, own forever. Around it sat five learning platforms — Learning Connector, LearningExchange, Precision LMS, Digital Guides, IoTU — five URLs, five logins, five lines on an invoice. Of 550K+ registered learners, only 350K+ ever came back. The official metric counted the gap and never asked why.

I spent three weeks inside the customer-success call recordings. Nothing was wrong with the navigation. Perpetual licenses meant no recurring revenue. No recurring revenue meant stale content. Stale content meant engineers learned on YouTube instead.

The brief said "redesign the UX." I argued the contract was the broken interface. The CRO had 60% of revenue sitting on perpetual licenses and said so loudly. The President of PTC University took the phased version: new customers on annual subscription from Q3 2017, existing ones protected for 24 months, then migrated.

Twelve months on: 64% of new bookings were subscription, against 0% in Q3 2016. The grandfathered cohort held at 78% through its first migration. Perpetual was 60% of total revenue when I started.

Five logins was the real onboarding flow.

An engineer certified on Precision LMS who wanted to learn on LearningExchange filed an IT ticket and waited three days for an account, then landed on an empty profile — no shared progress, no shared history, no shared recommendation graph. Wanting to learn more was the friction.

With four engineers and one PM, I shipped single sign-on across all five platforms in 14 weeks and rebuilt the IA as a graph, not a tree: a learner on Precision LMS now saw "engineers like you also learned on IoTU" without leaving the surface.

Cross-traffic became the biggest acquisition channel the under-used platforms ever had. IoTU and Digital Guides — PTC's IoT and AR bets for the next decade — got their first organic learner growth from that graph, not from the marketing site.

Nobody searches the way marketing writes.

The catalog mirrored the release calendar. Every page read "Precision LMS 5.0 — New Features." Product Marketing wrote it that way because that's how the business shipped. Engineers think in problems, not version numbers.

I argued for a use-case spine instead: "How do I run a tolerance analysis on a sheet-metal assembly?" rather than "Precision LMS 5.0 Features." I lost the review twice. I won the third with the search-query log from logged-out users — every query was problem-shaped, zero were feature-shaped.

I rewrote the content architecture against that spine. Enrollment rose 28% the quarter after launch, against the prior four-quarter rolling baseline.

I shipped the Netflix homepage. It was the wrong product.

The first unified landing screen showed all five product families to everyone, ordered alphabetically. I thought it would feel modern.

An engineer in Pune with a Precision-only license logged in and found 80% of the surface advertising platforms she had no right to open. Session length fell 19% in week one.

Three weeks later I keyed personalization off the license: the homepage now showed only what the customer's company had bought, with everything else behind one "Explore" link. Session length recovered, then ran 31% above the pre-launch baseline.

The user's contract is the most accurate filter you'll ever ship. I learned that by getting it wrong first.

A user guide with a shipping cost.

Every product shipped with a 200-page printed manual. Printing, freight, and warehousing them cost PTC roughly $1M a year. The hidden cost was time: a printed guide locks the product to a release cycle, because the next correction waits for the next print run.

I led the digital guidebook that replaced it — searchable, in-context, downloadable per chapter, updateable inside the platform. Print volume fell 92% in the first year. The $1M became annual recurring savings, against the 2016 print budget on a six-quarter rolling window.

The dollars weren't the real win. The guide stopped being a release artifact and became a living surface tied to the subscription. Once the contract changed, the guidebook could change with it.

Translation was an architecture decision.

I built the IA knowing German and Russian run 30% longer than English: short labels, shallow hierarchy, no text baked into images, 30% width buffers everywhere. Translators were testing in week 3, not week 20.

I added Portuguese (Brazil), Korean, and Russian for PTC's enterprise expansion — eleven locales, each WCAG AA, validated at 99% automated pass plus manual screen-reader and keyboard testing per language.

The side effect: designing for accessibility made every interaction fast on the 3G handsets engineers in emerging markets actually carried. Mobile went from 4% of sessions in 2017 to 38% in 2019, across all 80+ countries.

The receipt.

Every number with baseline + window

What it left me with.

Most platform redesigns stall because the design team isn't allowed to ask the contract question sitting underneath them. Every move here — subscription, SSO, use-case content, license-keyed personalization, the living guidebook, locale-first architecture — wore UX clothing over a business-model skeleton.

I run that check on every AI product now. When adoption stalls, "make it easier to find" is sometimes the answer — but "who pays for this, and does paying make them use it?" is what finds the real ceiling.

Design patterns demonstrated

• AI Failure States: graceful fallback when the adaptive recommendation system couldn't determine a learner's next course (low confidence, new profile, conflicting signals).
• Human-in-Loop Patterns: override mechanisms so learners could skip the system's recommendations and request a manual counselor review.