Case study · B2B SaaS
Designing a full product for a category that didn't have design patterns yet
How I designed a B2B advertising suppression platform from PRD to shipped product, brand included, in a novel category with no obvious references.
- Year
- 2024
- Role
- Product Designer (contract, sole designer)
- Industry
- B2B Advertising · SaaS
- Team
- PM-led, cross-functional
By 2024, B2B advertisers had a specific and expensive problem. They were spending real money on ads across Meta, TikTok, LinkedIn, and every other platform their customers used, but they had no way to stop showing ads to people who no longer needed to see them.
You bought the shoe on Tuesday. On Wednesday, you were still seeing the ad for the same shoe. The advertiser was still paying for that impression, and every impression like it, at scale, every day.
The problem wasn’t that ad platforms lacked audience targeting. That part worked. The problem was suppression: the ability to say don’t show this ad to that person any more, in this specific situation, across all of these platforms. Ad platforms didn’t solve this natively because their business model runs on impressions. So the work fell to the advertiser, and the advertiser’s tools were: manually maintained suppression lists uploaded to each platform separately, spreadsheets, and hope.
The insight
The platforms will never fix this, so the advertiser has to fix it themselves, from outside, with tooling that sits on top of all their integrations at once.
A suppression list uploaded on Monday was already out of date by Wednesday. Multiply that across every platform an advertiser used, every audience segment, every campaign, and the wasted spend was significant and largely invisible. My job as sole designer was to translate that insight into a product B2B marketers would actually use.
WasteNot was a contract engagement. The founders had already validated the problem and the PM had drafted a PRD. My role was to translate that PRD into a full product surface (dashboard, audience suppression rule builder, campaign management, platform integrations, analytics) plus the brand identity that carried across all of it.
Solo designer. Brand and product. Working to contract velocity, which meant no time for extended user research or testing rounds. Every design decision had to be right enough to ship, first pass, and hold up in production. Three things shaped the approach:
Design against a validated PRD, not from scratch
The founders knew the problem intimately (both had personally felt the pain of running ads that showed to already-converted customers). The PM had translated that into requirements. My job wasn't to rediscover the product. It was to shape how it looked, felt, and worked in the hands of users who'd never used anything like it before.
Design for a category with no reference patterns
Audience suppression across multiple ad platforms with rule-based dependency logic isn't a well-established design space. No obvious competitors to reference, no established interaction patterns to lean on. Every core surface had to be figured out from first principles: how to visualise a suppression rule, how to make dependency logic feel like a filter rather than a database query, how to show a marketer their ad spend is being protected without burying them in configuration.
Design one interface for three different technical fluencies
WasteNot's users ranged from sales-focused marketers who wanted guardrails and defaults, to marketing managers coordinating campaigns across teams, to data-driven operators comfortable with complex rule builders. One interface had to serve all three without dumbing itself down for the top of the spectrum or overwhelming the bottom.
That third constraint shaped every meaningful design decision on the product.
Before designing anything, I mapped WasteNot’s users into three distinct personas. Not because personas are a design ritual, but because the range between them was wide enough that ignoring it would have led to a product that served none of them well.
Sales-focused Sam
New to programmatic advertising, motivated by hitting quarterly targets, minimal appetite for setup complexity. Sam needs the platform to work with minimal configuration. Give him a rule builder with 20 conditional operators and he closes the tab.
Marketing Manager Maya
Coordinates campaigns across multiple team members and channels, needs oversight without micromanaging every setting. Maya cares about clarity, consistency, and the ability to hand a campaign off to someone else without them needing to be retrained.
Data-driven Dan
Comfortable with analytics, wants granular control, thrives on visualising data and finding patterns. Dan will happily configure complex suppression rules if the interface lets him. Hide the depth from Dan and he reverts to his manual workflow.
The design problem was that these three couldn’t be given three different products. One interface had to serve all three at once.
The pattern
Progressive disclosure with sensible defaults. Every surface had a simple, opinionated default state that got Sam and Maya productive fast, and an advanced layer Dan could expand into when he wanted control.
The defaults weren’t watered-down versions of the advanced setup. They were opinionated choices based on what most B2B advertisers actually needed, so a Sam-level user could ship a real campaign without ever touching the advanced controls. That framing carried through every core surface: dashboard, audience library, campaign builder, suppression rules, integrations, and analytics.
WasteNot shipped as a full B2B advertising platform with six core product surfaces, plus a brand identity that carried across all of them. Live in the US, integrated with the ad platforms B2B marketers were already using.
The product hung on two flagship flows: connecting your ad stack, and building the suppression rules that protect your spend. A calm dashboard summarised the state of both, with analytics, campaign management and the brand system wrapped around them.
The integrations manager handled the platform connections to Meta, TikTok, LinkedIn, and the rest. Each was designed to be set up once and forgotten; status indicators told users at a glance whether a connection was healthy, whether data was syncing, whether attention was needed. The surface was invisible when it worked and legible when it didn’t.
The core design challenge
The suppression rule builder is where a marketer says “don’t show this ad to anyone who has seen it in the past 2 days, unless they’ve clicked through and haven’t converted yet.” A preset default let Sam ship a campaign in under a minute. The advanced view let Dan build whatever he needed.
Campaign management let Maya coordinate work across her team, every campaign visible in one view. Intuitive checkboxes made selecting the right groups and campaigns fast, and duplication was a first-class action, because in B2B advertising most new campaigns are variants of existing ones.
Audience management consolidated settings and platform connections across every connected ad platform into one place. Before WasteNot, a B2B marketer had to log into Meta, TikTok, and LinkedIn separately; here, everything sat together, with a clear indication of where each audience came from and when it was last synced.
Around those two flows sat the rest of the product. The dashboard was the first thing Sam or Maya saw when they logged in, not a wall of metrics but a calm summary of what mattered right now: which campaigns were live, which platforms were connected, how much wasted spend had been prevented this week. Analytics and reporting gave every user the same underlying data at different depths, from Sam’s single headline number to Dan’s per-audience, platform-by-platform view. And the brand identity was designed alongside the product, not after it, so the marketing surface and the product surface felt like the same object.
WasteNot was a different shape of design work from most of what I’ve done. Not a 0-to-1 fintech with months of research and iteration. Not a design system embedded inside a large product organisation. A contract engagement, shipped fast, designing against a validated PRD in a novel category, with no time for the extended user testing that would have made me more comfortable with every decision. Working that way taught me a few things I’ve carried into every project since.
Design intuition is a real tool when the timeline demands it
Contract velocity meant I couldn't test the suppression rule builder with real users before shipping. What I could do was interrogate the PRD carefully, understand the three user types intimately, and design with a bias toward defaults that were opinionated but forgiving. That's a different discipline from research-led design, and a legitimate one when the constraints are honest about themselves.
Progressive disclosure is a genuine answer to the design-for-range problem
Most B2B products fail one of two ways: they overwhelm beginners with configuration, or they patronise power users with a locked-down interface. Sensible defaults with an advanced layer available on demand is a pattern that scales across technical fluency. WasteNot proved it out in practice.
Brand and product should be designed as one thing
Because I owned both, I got to see how much stronger a product feels when its brand identity and its interface come from the same design system. Not the same visual system by accident, but the same underlying design decisions, applied consistently from logo to microcopy to component. Most companies split these responsibilities across two designers or two agencies, and the seam shows.
The most valuable thing WasteNot taught me is that novel categories reward first-principles thinking more than pattern-matching. When there’s no obvious reference product to copy, you’re forced to actually understand the user’s problem and design a real answer to it. That constraint isn’t a limitation. It’s the condition under which the strongest design work usually gets done.
“Novel categories reward first-principles thinking. When there's no reference, you're forced to design a real answer.”
Design principle carried into all subsequent product work
Building a Design System for an AI-First Team
