How Lucen Software tied ad spend to ARR and improved product activation from ads by 167%

When Eddie Casado joined Lucen Software as Performance Lead in July 2026, Google Ads said Lucen's campaigns were earning back six times what they cost. Lucen's revenue didn't support that number, so Eddie connected Supermetrics and Lucen's revenue data in Claude Code to check it. Four months later he runs his reporting and campaign builds from there, and the share of new Lucen Track users who activate has gone from 2.4% to 6.4%.

Their challenge

Lucen Software makes Microsoft-native add-ins that help teams plan, track and communicate complex work in the apps they already use. Lucen Timeline turns project data into timelines and Gantt charts in PowerPoint and keeps them in sync with project management tools. Lucen Track lets teams track time and invoice clients without leaving Outlook. Lucen Flowchart turns processes into clear flowcharts in PowerPoint, and with Lucen AI it can build one from a prompt. Trusted by the majority of the Fortune 500, the company was called Office Timeline until May 2026, when a rebrand changed its name, website address and product design at the same time.

Paid ads drive most of Lucen's funnel. Eddie runs Google Ads, Lucen's biggest channel, plus Microsoft Advertising, LinkedIn and a ChatGPT ads pilot, and tests ad messages on Reddit.

Before he arrived, performance reports were built in Excel, with ad platform data isolated from the rest of the martech and business tech stack. The reports showed clicks, costs and trends, but not which ads led to paying customers. Google Ads claimed a return on ad spend (ROAS) of about 6x and Microsoft Ads about 3x. Lucen was making money, but Eddie couldn't connect those ad conversions to a single customer. "Otherwise, how would we set up any experimentation and optimizations?" he says.

What they did

Lucen already had a Supermetrics license, but it wasn't being used much. Eddie had started comparing data in another tool but it wouldn't connect with the universe of tools he needed to contextualize performance well. So he moved the work into Claude Code and connected Supermetrics through the Supermetrics MCP server. That gave Claude his ad platforms and Google Search Console in one place. Next he hooked up Chargebee, Lucen's billing system, through code, and HubSpot and Amplitude came after that.

He wanted to know whether the 6x figure was real. Each new data source answered part of that question, and depending on the platform, the numbers told a different story. Over time the setup turned into Lucen's own system for working out which marketing brings in revenue.

How Supermetrics and Claude Code power it

1. Matching ad results to real revenue

The first big finding came from matching the sign-ups Google Ads reported with the contacts in HubSpot. The rebrand had broken parts of Lucen's tracking: the tags that record which campaign a visitor came from were being lost along the way, including on the pricing page. Google Ads also wasn't getting sign-up data straight from Lucen's own systems, wasn't told what each sign-up was actually worth, and couldn't tell work email addresses from personal ones, so it had no way to favor sign-ups from businesses.

Eddie then built reports showing how much Lucen spent on ads, how much revenue came back and how much of that revenue started with a campaign. He modeled the data with first-touch attribution first, then with last-touch. Most of the revenue was credited to something other than ads, partly because of how Lucen gets bought: an employee tries the product, then their manager pays for it. So Eddie matched the company email domains of people who signed up from ads with purchases from the same company in the same week or month. That turned up revenue the campaigns had likely helped bring in but weren't getting credit for. He then turned all of it into an attribution model in a Claude artifact, built on the ad platform data he'd brought together through Supermetrics.

Lucen has since added RudderStack to collect customer data more reliably. The revenue reports run on data from Supermetrics, Chargebee and Amplitude, and an operations colleague keeps them up to date along with a marketing dashboard built from Supermetrics.

2. A week in Claude Code

Eddie's day starts with a briefing that Claude puts together on a schedule. It reads his Slack conversations, call notes from Fathom and messages from colleagues, then flags the tests he needs to check and anything that finishes that day.

Every change he makes to a campaign is written to a log in Notion automatically, by platform and by campaign.

Other routines run on their own schedule:

• A weekly report on the searches that triggered Lucen's ads across Google Ads and Microsoft Advertising, with Supermetrics pulling the data.
• Dashboards that pull campaign data through Supermetrics and explain this week's results in light of last week's changes.
• A weekly scorecard for Lucen Timeline's new conversion tracking and bidding setup.
• An alert when a campaign has settled after a big change, so Eddie knows when to check it again.
• A check on how often Lucen's ads appear for the searches it's targeting, the number Eddie watches most closely in Google Ads.
• A daily check that Lucen Track's sign-up tracking is working.

3. From keyword research to a campaign ready for review

For a recent keyword project across Lucen's two Google Ads accounts, Claude went back through campaigns from before and after the rebrand, found longer, more specific search terms worth targeting, and matched them to the latest customer profiles in product marketing's Notion pages. Eddie adds what customers say on Reddit and LinkedIn, collected with Apify, and feedback from Canny and Enterpret. A Webflow skill built by Lucen's designer and web manager creates the landing pages in the site's existing design. Then Supermetrics creates the ads and ad groups from Claude's recommendations.

Work that took two or three days, sometimes a week, comes back in about two hours, so Eddie can do the same for all three products.

4. Reviewing changes before they go live

Claude builds each campaign from the personas and messaging product marketing has already written, so product marketing reviews a finished campaign before it launches. The Notion change log shows anyone what changed and when. Eddie still looks at new campaigns in the ad platforms once or twice a week, mostly to make sure there aren't any integrity issues with the campaigns, and also as a healthy habit to keep a closer pulse on campaigns. He estimates about 80% of his campaign work now happens through Claude, with a person approving every decision.

Outcome and impact

Eddie estimates he spends about 80% less time on manual work, and that projects which used to take three months now take about a week. He has put that time into fixing Lucen's funnel after the rebrand and testing new ad formats.

• Eddie used Supermetrics to match each record in Lucen's internal ledger with ad platform data by click ID. That showed him which campaigns brought in people who went on to use Track, as well as the ones that only produced sign-ups. With that view, he could choose which optimizations to make first. Lucen Track activation rose from 2.4% in August to 6.4% in October. That is a 167% improvement on product activation via advertising.
• Lucen has recovered at least 1% in its free-to-paid funnel since July.
• Putting ad spend, platform results and revenue side by side revealed a path to purchase that nobody had mapped. Product marketing has now mapped it, and the team is fixing the tracking gaps it found.
• A new experiment to get Lucen Track's sales team more demo requests is growing the sales pipeline and improving lead quality. Claude checks open deals against how long Lucen's deals usually take to close and tells Eddie when to wait before changing anything.

What's next

Eddie wants to close the gap between what Lucen measures today and what it doesn't measure yet, so the team gets the best possible read on campaign performance, drawing on all the data Supermetrics pulls together. As Lucen adds channels, he plans to add marketing mix modeling to measure what each one contributes. He has also built a competitor tracker in Claude Code. It watches competitors' websites, pricing, comparison and help pages, the keywords they bid on and how their ads compete with Lucen's, then posts red, yellow or green alerts to a Slack channel the company uses.

His advice for other SaaS marketers who want to run paid media from Claude Code: "Don't be scared of Claude Code. Make it work for you. You don't have to follow anybody's framework. Know what you hate, know what you love. I hate spreadsheets, so: build me a dashboard."