From Concept to Production in 5 Months

The business case was clear: retain and win high-value advertisers who wanted influencer and affiliate in one platform, grow the creator network by attracting influencers to the ecosystem, and open new revenue streams from creator-focused campaigns.

The challenge? Rakuten had never worked closely with influencers before. There was no infrastructure to connect social accounts, no management tools for campaigns, and no creator-facing UI. An added pressure was to meet a hard deadline for a prospect demo the following Monday.

Challenge

This project had many constraints that can make design hard:

• Extreme time pressure — a prospect demo within a week of project initiation
• New user type — influencers think about content creation, brand partnerships and social engagement, not metrics and commission structures
• New product category — campaigns were a new concept; advertisers would create opportunities, set deliverables and offer incentives, then invite and manage influencers
• Technical unknowns — we were using a partner to authenticate social accounts and consolidate posts linked to campaigns
• No existing patterns — although some influencers were in the platform, there was no influencer-specific UI; everything needed designing from scratch
• Cross-functional complexity — working across timezones, coordinating with external partners, aligning with developers estimating effort on the fly

My approach: Prototype first, validate fast

On Tuesday morning, I joined a briefing call with the VP of Product. The big question was asked: "Can we demonstrate a compelling vision of this product to a prospect by Monday?" That gave me less than a week.

I started with a light level of research to enable quick user flows, which were beefed out with wireframes and quickly iterated into a golden-path clickthrough prototype with high-fidelity mocks.

In five days I built a full clickable prototype covering the end-to-end experience: campaign creation, influencer invites, and application review on the advertiser side; campaign discovery, application flow, and deliverable submission on the influencer side; social post display; and all key workflows connected and demonstrable.

Jumping straight to high-fidelity mockups — using existing design system components where possible — meant the prototype moved fast without sacrificing credibility. Prototyping the complete flow, rather than individual screens, surfaced technical questions early and gave engineering clear requirements to evaluate. When the VP of Product saw it, she had enough confidence to demo the vision to the prospect directly. Prototype delivered Friday, demo on Monday, project greenlit for December MVP.

Key Decisions

With the prototype validated, I moved into detailed design and scoping. Here are the critical decisions that helped shape the product.

Ruthless V1 scoping: ship a complete story, not everything

With only five months to production, I worked with the PM to make hard choices about what to include in MVP versus what to defer.

Everything went through the same filter: does this need to exist on day one?

If a user could complete that journey end-to-end, we had an MVP. The ruthless scoping had an unexpected benefit — it forced me to identify the absolute core value proposition. What's the one thing this product must do? Let advertisers create campaigns and let influencers apply and complete them. Everything else is enhancement.

Dual-audience UI

Advertisers and influencers needed fundamentally different experiences from the same underlying campaign system.

Advertisers needed business-focused tools: a management dashboard tracking applications and completions, detailed influencer profiles with audience data and engagement rates, and language around goals, deliverables, and compensation structures. Influencers needed almost the opposite — opportunity-focused framing that answered "what's in it for me?" before anything else. Campaigns needed to be sold, not specified. The application flow had to be quick and low-friction, with requirements framed as expectations rather than obligations.

These couldn't be separate products (engineering would never deliver in time), but they couldn't feel identical either. I settled on shared underlying components and data structures, but with different content hierarchy and visual framing. Advertisers see "Campaign Management." Influencers see "Opportunities."

Social post verification

The most critical — and most uncertain — piece was social post detection. The external partner would detect when an influencer posted about a campaign, but the exact mechanism wasn't finalised. How do you design a verification UI when you don't know exactly what data you'll receive?

• Design for the ideal state first — what should it look like if everything works perfectly? Show the post preview, engagement metrics, verification status
• Then design for failure states — what if the post isn't detected? How does an advertiser manually review or flag issues?
• Build in flexibility — card-based layouts that can accommodate variable data, whether posts have images, videos, or just text

Influencer profile design

Advertisers needed to evaluate influencers before approving applications. We suddenly had access to social data we'd never dealt with — follower counts, engagement rates, post frequency, audience demographics. It would be easy to show everything, but more data isn't always better. I prioritised active socials, engagement quality over vanity metrics, recent activity, and audience relevance. Everything else was cut to keep the profile clean and the story immediate.

Campaign details page

After clicking a campaign thumbnail, influencers needed comprehensive information presented persuasively. I shifted my thinking from "display campaign details" to "convert interested influencers into applicants." Key decisions: lead with deliverables and compensation; frame requirements as opportunities, not obligations; clear primary CTAs with minimal friction. The page needed to feel like an opportunity worth pursuing, not a dry specification document.

Testing and iteration

Solution

The final MVP design focused on the core workflow: advertisers create campaigns, influencers discover and apply, content gets posted and verified, incentives get distributed. Key design elements: progressive disclosure to show essential info first; card-based layouts flexible enough to handle variable data from external APIs; clear primary CTAs on every screen; dual-audience framing with different language and hierarchy for each user type; and visual hierarchy that makes incentives and requirements unmissable.

Outcome

• Complete clickable prototype — 5 days
• Prospect demo and buy-in — Day 6
• MVP development — September 2025 to February 2026 (5 months concept-to-production)
• Planned V2 iteration — Mid-January 2026 based on real user feedback and data

Moving fast with a tangible prototype did more than hit a deadline. It secured the prospect's commitment before competitors could respond, validated the concept before any serious engineering investment, and aligned stakeholders around a shared vision they could actually click through. The prototype also created its own momentum — when people can see and interact with something real, it stays prioritised through competing demands in a way that a slide deck never does.

What I learned

Having a prototype that looked and felt real made a concrete difference. A VP could show it to a prospect with confidence — wireframes wouldn't have done that. Fidelity bought credibility faster than I expected.

The five-month deadline forced a kind of clarity I rarely have on longer projects. Every feature had to earn its place in V1. I ended up cutting things I would normally have kept, and the product was tighter for it.

I built the layouts to be flexible from the start — cards that could handle variable data, components that could be extended without breaking. That wasn't over-engineering; it was the only practical way to ship something that could grow once real users showed up with real feedback.

Prototyping end-to-end, rather than screen by screen, helped me catch gaps that would have been expensive to fix later. It's easy to design a screen that looks good in isolation. It's harder to design a journey that actually holds together.

And the deadline, oddly, helped. When you have to ship something complete in a tight timeframe, you stop defending nice-to-haves and focus on what the product actually needs to be usable.

Other case studies

Natural Language Search & AI

Multi-Touch Attribution for Affiliate

Enhancing offer management