Works.

Scandit's barcode scanning was a category leader. ID scanning was adjacent in tech but very different in buyer, use case, and competition. We recognized two very different persona categories: developers / technical buyers, and industry operations.

For the dev/tech persona we stayed clear and to the point – capabilities, tech differentiation, integration options, simplicity.

For the industry ops persona we focused on the jobs-to-be-done of each industry and use case – airlines, logistics, post & parcel, retail. We mapped the problems we solve in each industry and the real workflows we improve.

Positioning by persona, vertical messaging, sales enablement, launch video, web pages, demo experience.

Scandit's growth had been SDK-to-developers. The strategic shift was to lead with industries and use cases instead – retail, post & parcel, logistics, field service, air travel, healthcare. Each backed by a real customer JTBD.

Researched each vertical with product, sales, and customers. Translated findings into tailored positioning and messaging per industry. Built use cases that the buyer recognised in their own workflow.

Then partnered with content, design, and sales to turn each vertical's positioning into a full asset stack: web pages, whitepapers, videos, pitch decks, sales enablement.

Vertical web pages and messaging across retail, post & parcel, logistics, field service, air travel, healthcare. Whitepapers (incl. retail's "Think Like Amazon"). Industry "dollhouse" visuals mapping use cases across operations. Vertical pitch decks. Sales enablement per industry.

As a Samsung solution partner, we needed content Samsung's reps would actually use. Both teams wanted to show that barcode scanners are old and clunky for retailers – and shift perception toward smartphones. The risk: the negative framing would stick to our own brands too.

Found a visual and narrative reference that flipped the frame: retro. Instead of calling smartphone scanners "old", we called them retro – memorable, slightly humorous, no insult. The retailer doesn't have to defend their past choice to consider a new one. Packed the video with real examples where smartphones beat scanners.

Co-produced launch video, split-budget production, joint distribution through Samsung sales channels.

ShelfView was a completely new product line: shelf computer vision intelligence for retailers. We had to build trust, beat competitors on visual quality, and educate the market on a category Scandit hadn't been known for – in a short runtime.

Narrowed the value prop to the few elements that resonated. Mixed three modes: technical animation (the how), retail store scenes (the problem), and on-camera presence from two Scandit team members and one customer (solution and trust). Skipped voice-over – let the people speak. The mix made it engaging without diluting the message.

Launch video, supporting positioning, web content, sales materials for the new category.

I produced 100+ product videos. Early on, one came back from the shoot, the CPO watched the final cut, and said he didn't like the intro.

Video is opinionated, and there are a lot of stakeholders. If you don't manage that, a project would turn into a committee that waters everything down.

I built a staged sign-off system – a pyramid. Each stage got signed off before it fed the next: brief & goal → positioning → key idea → narrative flow → storyboard → final script → final video. A RACI table made it clear who approves and who's just informed and doesn't need to weigh in.

So when someone asked late "why aren't we saying benefit X in the intro?", I pointed back to the positioning we'd already agreed. It killed group conformity and late rewrites – and let us ship punchy, strong videos instead of safe mush.

100+ product videos across launches, web, sales decks, events, and analyst briefings. A repeatable staged process and RACI other teams reused.

Selling a computer vision SDK to developers breaks the usual playbook. Most things you can create demand for – show someone the right ad and they might buy. An SDK isn't like that. A developer reaches for one only when they hit a real need, runs a quick test, and moves on. And computer-vision quality is notoriously hard to understand – you can't just claim performance, you have to prove it. The job is to be useful, and technically honest, at the few moments they're paying attention.

I built PMM around the moments they care about – evaluating, integrating, and shipping into the real world. Four moves, each one earning trust by being useful and technically honest...

Devs assume "if it scans on my desk, it works." It doesn't. I made the real-world gap impossible to ignore. See the project →

No adjectives. Measured, side-by-side proof against the exact tools devs were comparing us to. See the project →

A complex SDK, tangled integration paths, and dense releases – made clear in seconds, in a look devs respect. See the project →

Hands-on demo apps and books, so devs feel the product in their own setting before writing a line of code. See the project →

As Scandit moved from an SDK to offering ready-built UX data-capture solutions, we hit a mindset problem: developers thought they could build the capture experience themselves – and did, but without real-world UX testing. From the desk it worked – job done. In the field the real user experience was bad, and developers never found out.

These teams don't run real usage tests, so the gap was invisible to them. It had to be felt. I sat with the UX team, pinned down the key failure areas, and built a video that mixed real human reactions, split-screen comparisons, and a clear UX teardown – simple to follow, informational, but with emotion.

An educational video on why scanning UX matters – human reactions, split-screen comparisons, and a plain-language explanation, plus Scandit's ready-to-use UI and no-code path to getting it right without rebuilding from scratch.

We had to make it land across a range of personas – IT, operations – not just software engineers. So it explains the stakes without assuming you write code.

Computer-vision performance is notoriously hard to understand. In a quick, clean test almost anything scans – so a free or "good enough" option can look equal to ours. Many developers were defaulting to free, open-source scanning libraries for exactly this reason: in a simple test they looked the same. The real difference only shows under pressure: scanning at sharp angles, through glare, in motion, the false positives. Numbers on a slide don't land, and most buyers never test those conditions themselves.

So I teamed up with product engineering to pin down exactly where we win – the specific, real-world conditions that break a weaker scanner. We even built tools to measure and visualise things like scanning distance on video, then ran controlled tests to quantify the gap.

Benchmark report. A head-to-head performance comparison against three key competitors – packaged into a report and a clean set of benchmark numbers sales could drop straight into a deal. Hard data on where we win, in the conditions that decide real deployments.

Side-by-side comparison video. We filmed our scanner head-to-head with the competitors in the same hard conditions, with the test results on screen – so the gap was visible, not just claimed.

Performance explainer video. A video that breaks down the categories of computer-vision performance and shows our CV in action across each one.

Early on, release comms were written in marketing language that developers tuned straight out. Our integration options had grown complex and hard to understand. And releases were big and feature-dense.

Release comms, to the point. On one of my early projects we learned marketing language didn't work here at all – so we restructured the copywriting: written by the product manager with PMM assistance, not by a marketer as before.

Integration options, made pickable. I built web pages that sorted the options by their real decision points – a matrix that made it simple to choose the right one. It took clear, simple technical thinking to present something that complex this plainly. An animated explainer, in a high-end technical look devs appreciated, backed it up and gave Scandit the feel of the best tech solution.

Bento-box release visual (SDK 7.0). For our big SDK 7.0 release I used a bento-box visual – Apple-style – as the hero for all the key new features, in animation and static visuals. It landed, got reused across many later releases, and spread into marketing materials.

Scandit's software comes alive in the physical world – scanning real labels, AR overlays on real shelves. We learned very quickly that selling the technology and its features wasn't enough. Each persona – retail associate, field-service tech, logistics worker – needed to see their own workflow, exactly as they experience it, to connect the dots and imagine it in their facility.

So we stopped demoing "the technology" and started demoing the buyer's day. We designed every experience around each industry's real tasks and the exact information a worker sees on the job – before they ever tested it in their own facility.

Interactive trade-show demos – walls you point at to scan products and see AR relevant to your industry. Among the most interactive demos on the show floor.

Demo books – let prospects capture their own setting, then scan, trigger AR, and see the info a worker sees, before testing it in their facility.

Demo apps – built closely with the PM, planning the best app experiences to show our capabilities, down to exactly what product info to surface for workers in retail, field service, and logistics.

Indoor-location IoT technology that could work across many industries. As a small startup out of Dublin, we needed focus, trust, and attention – without the size to command any of it.

We started in healthcare, but the sales cycles were brutally long. So we switched to logistics. Then I built a GTM engine that made us read like a much bigger company than we were.

Started content and SEO very early – which paid off fast. Learned PR hands-on: small wins in Irish tech news first, then features in key logistics and IoT magazines whenever we had something to announce – Inbound Logistics, Supply Chain Digital, Supply & Demand Chain Executive, IoT Now.

Built an outbound system on the CRM and marketing automation. Wrote the value proposition, whitepapers, and ROI calculations. Pulled it all into a website and content engine.

From talking to language learners, we recognized two key problem patterns. The gamified apps trained people to tap: streaks, hearts, gems, dopamine… and years in, they'd still freeze at the bakery. The story apps rely on reading longer texts – but focused practice is deprioritized, so people learn mostly passively.

Both types don't offer efficient learning for the time a learner spends on the phone – just the feel of learning.

So we built the opposite. Learning in context, but with micro stories. The premise: you don't need much text to get the most out of learning. It's methodical but guided – a calm, minimalist experience.

The stories come from real news, written one CEFR level at a time through a multi-step, AI-assisted process – so they're interesting, factual, and use the right language. None of the AI slop and low-quality stories you find elsewhere.

Every story runs the same four steps. Read for comprehension. Meet the vocab. Pick up grammar phrases. Then write it back in your own words, with AI correcting you as you go.

We built unique list-based flashcards – nobody else has this on the market. And the spaced repetition is visualised simply.

Product direction, UX, positioning, the AI content pipeline, the German story system, website, launch and GTM. Built in a team of three.

170+ stories live across A1–B2.

See it live: upwordo.com · open the app.

Embedded recruiting was proven abroad but unknown in Switzerland. A new venture needed the whole stack – category story, name, brand, site – built fast and lean.

Research, positioning, naming, copy, design and the build all ran through one AI workflow I created – then handed off, so the team keeps producing without me.

Named the third option – between an agency and a full-time hire – and kept the buyer's pain in view throughout. The category makes it click; the brand stands on what's harder to copy than a label.

Recruiter (Within) – memorable, ownable, and you read what it means at a glance. The parentheses became the brand mark.

A longer read that teaches the model, at recruiterwithin.com.

Email campaigns going out cold to industry leaders – HR, finance, pharma, and more. Conversion stuck. Buyers ignored generic webinar pitches.

Switched from "register for this webinar" to giving the value upfront. For each webinar, I worked with the speaker to put useful information right in the email – so people could experience the value before deciding to show up. Value first, ask second.

Content strategy, 100+ campaigns, blog launch, infographics, webinar landing pages.

Niche identity & access management software (midPoint, open source). Deep technical content, low organic traffic. The team's instinct was to write more, faster.

Audited what was already there. Their product wiki at wiki.evolveum.com sat on a subdomain and had quietly built strong domain authority over years – open-source IAM is a rare, deep niche, and that wiki had become a reference. But it barely linked back to the main site. The asset was already there. Use it.

Mapped the highest-intent keywords to pages on evolveum.com. Built internal linking from the wiki into those pages – the authority flowed where it could convert. Layered the rest on top: URL and meta restructure, Wikipedia references, Quora and StackOverflow answers, SlideShare decks repurposing existing content, technical PR.

Keyword strategy mapped to pages, wiki-to-main-site internal linking, URL and meta restructure, breadcrumbs, off-page program (Wikipedia, Quora, StackOverflow, SlideShare, technical media).