Work · Chapter 02 · Canonical

A cost centre that learned to invoice. Social went from unmeasured to $5M+ in pipeline influenced.

Communications Manager · 2022 to 2025 · 2 min read

Canonical is the enterprise open source company behind Ubuntu. Twenty plus products, one feed, and a question that kept coming back: what does social actually earn?

Results at a glance

$5M+

pipeline influenced

One year

36 → 785

social-influenced MQLs

2022 to 2023

4,731

Data and AI Masters registrations

One event series

The stakes

Social had no link to leads.

20+ products fighting for one feed, and no line from a post to a lead. Which means social was a cost centre with good engagement, and nobody could say what it earned. That isn’t a reporting problem. It’s a structural one, and it doesn’t fix itself by posting harder.

My part

What I was actually responsible for.

RoleCommunications Manager
Period2022 to 2025
DesignedThe UTM and campaign tracking, linked to Salesforce by marketing operations

The moves

What I ran.

01

A calendar by topic bubble (AI, cloud, security, automotive), weekly then fortnightly, with partner series: AI Summer Camp, Data and AI Masters, HPE and HP, AWS GovCloud. One feed serving twenty products without becoming a wall of logos.

AI Summer Camp post on LinkedIn
AI Summer Camp post on LinkedIn

02

Every webinar rerun and cut into 1-minute trailers and clips, because the recording is the cheapest content you will ever make and most teams never go back and touch it.

Automotive trailer: secured open source Linux for EV charging
Automotive trailer: secured open source Linux for EV charging

03

A scoring system on quality, engagement and conversion, with MQLs delivered to sales and marketing, organic and paid, which is the difference between reporting on social and being accountable for it.

The proof

Social leads went from 36 to 785.

Social-influenced MQLs

3620227852023

Tag, route, feed, report

  1. 1 Tag: UTM and campaign tracking on every post
  2. 2 Route: tracking linked to Salesforce
  3. 3 Feed: webinars, whitepapers, lead forms
  4. 4 Report: OKRs and C-suite QBR data

The blog

The canonical.com blog, captured 17 November 2026.
The canonical.com blog, captured 17 November 2026.

Then · 2020

An archived capture of canonical.com from 2020.
An archived capture of canonical.com from 2020.

In their words

Thanks a lot for all the work you’ve done to raise the bar for social at Canonical and for putting Canonical on the virtual ‘great companies of this world’ map.

Thibaut Rouffineau · VP of Marketing, Canonical

His strategic direction played a key role in driving significant growth across Canonical’s social media platforms.

May Myo Set Htwe · Social Media Management, Canonical

Want this for your company?

In the wild

The work, in the wild.

Automotive trailer: secured open source Linux for EV chargingLoads YouTube when you press play. Captions on.
Webinar: AI at the edge
Webinar: AI at the edge
LinkedIn Live: Data and AI Masters
LinkedIn Live: Data and AI Masters

What it means

So I built a profitable demand generation machine out of live and scheduled webinars and video streams. The tracking came first: a UTM and campaign structure designed so that a social-influenced MQL could actually be counted in Salesforce. Which raised the obvious question: what would happen if we asked the same feed to produce demand instead of impressions? 36 to 785 in a year, and $5M+ in pipeline influenced.

The long read

If you build it, they will come. If you can’t measure it, they’ll ask.

Social at a B2B company has a credibility problem, and it isn’t the reach. It’s that when the CFO asks what social did last quarter, the honest answer used to be a shrug with a very nice engagement rate attached. Field of Dreams worked because the audience showed up. It doesn’t work when the board wants to know what they did after they showed up.

Canonical had five portfolios, twenty-plus products, and a social operation that generated genuine attention and no provable revenue. That’s not a marketing problem. That’s an accounting one. So the first thing I built wasn’t content, it was plumbing.

A UTM and campaign structure designed so a social-influenced MQL could be counted rather than guessed. Every topic bubble, AI, cloud, security, automotive, got its own calendar, weekly then fortnightly, with partner series. Every webinar got rerun and cut into one-minute trailers, because a recording is the cheapest content you will ever own.

The scoreboard nobody argues with

Then the interesting part, and the part I’d defend hardest. Once you can sew a lead to a post, social stops being a cost centre and starts being a channel with a P and L. 36 social-influenced MQLs in a year became 785. $5M+ in pipeline influenced. 4,731 registrations for one event series, Data and AI Masters, with no paid reach behind it.

I’d love to claim that came from better posts. It didn’t. It came from a scoring system on quality, engagement and conversion, and from handing MQLs to sales instead of handing a dashboard to nobody. The content was already good. The measurement is what made it count.

What it actually changed

There’s a version of this story that’s about social media, and it’s the boring version. The real one is about what happens to a team that can finally prove its own contribution. People stop asking permission and start making bets.

If I had to argue for one thing, it’d be this: attribution is the least glamorous job in marketing and the one that decides whether you exist in two years. It’s the sound department on a film. Nobody walks out talking about it. Take it away and the whole thing falls apart.

So if I were starting this again tomorrow, what would I build first? Not the calendar. Not the content. THE PLUMBING. Everything else is decoration until you can prove something.

Canonical culture post on LinkedIn

Next story · Chapter 02 · Canonical

560,395 followers. Second only to Red Hat.

560,395

LinkedIn followers, up from about 90K

Your company could be next

Talk about doing this for your team.