CASE STUDY · DROPBOX · SaaS PRICE INCREASE

Making a price increase feel like
a helping hand

When Dropbox migrated legacy business plans to new pricing tiers, every admin had to actively choose. The content challenge: reduce cognitive load, earn trust, and protect revenue — all in a four-screen flow.

Content Design

\\\

UX Writing

\\\

Growth

\\\

Pricing

\\\

Monetization

\\\

Content Design \\\ UX Writing \\\ Growth \\\ Pricing \\\ Monetization \\\

My Role
Content Designer

Audience
Business IT admins managing team plans

Team
Product Design, PM, Data Science

Key Outcomes
Positive trends across flow completion, plan selection, churn & refunds

Surface
Web — Admin dashboard

The problem

Dropbox was transitioning legacy business plans to new pricing tiers — Standard and Advanced — at a higher price point. This was before subscription price increases became normalised in the industry, so the stakes were high: every team admin would need to actively choose their new plan.

Any friction in that flow risked churn, refund requests, and lost revenue. The content design challenge was twofold: how do you explain a complex pricing change clearly enough that busy IT admins don't abandon the flow, and how do you frame the options so the "right" choice feels obvious — without being manipulative?

What I owned

Specific Contributions
I was the sole content designer end-to-end — from auditing the decision landscape and identifying an unexamined default recommendation strategy, through writing all copy across the four-screen migration flow, to ensuring the language served both the decisive admin and the skeptical one.

I worked alongside product design, product management, and data science throughout. My content framing directly shaped a key product decision: the recommendation of the Advanced plan as a default rather than presenting both plans with equal weight.

The outcome

A seamless self-serve migration experience that guided admins to a plan decision — and kept them there.

↑

Flow completion

↑

Advanced plan selection •

↓

Churn

↓

Refund requests •

↑ Flow completion ↑ Advanced plan selection • ↓ Churn ↓ Refund requests •

All four key metrics tracked by the data science team trended positively. The default recommendation strategy — grounded in a genuine usage-fit rationale rather than a dark pattern — proved that reducing decision load and earning trust are complementary, not in tension.

Screen recording of the final flow. The experience begins on the landing page presenting Business account team admins with a review of the upcoming transition and recommended plan. The admin can choose to continue with the preselected option or explore alternatives.

  • Auditing options · Identifying a third path · Writing for two emotional states

    The work unfolded in four stages, each one informing the next:

    01 Auditing the decision landscape

    I mapped the options already under consideration — an equal-weight plan selector, and a questionnaire-based helper flow — evaluating each across two axes: how directional vs. open-ended, and how many steps were required. Neither was clearly right.

    02 Identifying a third path

    Through this audit, I surfaced an option the team hadn't considered: recommending the Advanced plan by default. Most legacy Business users already had features that mapped to Advanced — so pre-selecting it removed an unnecessary decision point while maintaining admin agency to switch. This reframe shifted the product direction.

    03 Designing the four-screen flow

    The final flow — Set Context → Inform User → Allow Alternative Selection → Confirm — required copy that was transparent about the price change, specific about plan differences, and warm enough to defuse frustration without over-explaining. Each screen had to earn trust in sequence.

    04 Writing for two emotional states

    Some admins would arrive informed and ready; others would feel ambushed. The copy had to serve both — efficient and scannable for the decisive admin, reassuring and thorough for the skeptical one — without creating two separate flows.

  • The line between helpful defaults and biased patterns

    The most consequential decision was also the most contentious: recommending Advanced by default. The concern was legitimate — defaulting users toward a higher-cost plan in the middle of a price increase could easily read as manipulative. The content had to do a lot of work to make the recommendation feel grounded in the user's actual situation, not Dropbox's revenue goal.

    The default wasn't about hiding the cheaper option — it was about not making admins do work they didn't need to do.

    I anchored the recommendation in usage continuity: the message wasn't "choose Advanced," it was "your team already uses these features — this plan keeps things as they are." That framing made the default feel like a service, not a nudge. The Standard option remained fully visible and accessible throughout.

  • Content design as product strategy · Reducing decisions, not options


    Content design is decision design.     The most impactful contribution on this project wasn't the copy — it was identifying a product direction the team hadn't considered. Auditing the decision landscape and reframing the default was content strategy at the product level.

    Reducing decisions ≠ reducing agency. Recommending a default while keeping the alternative fully accessible proved that you can reduce cognitive load without reducing choice. The admin still had full control; they just didn't have to do unnecessary work to reach the right answer.

    Trust is sequential. Each screen in the flow had to earn the right to the next one. A price increase announcement that feels ambushing on screen one makes every subsequent screen harder to read charitably — so the first screen's tone set the ceiling for the rest.

  • How this project could be reimagined with A.I. today

    From a smart default to a genuinely personalized migration

    The Morpheus default recommendation worked because it was grounded in a reasonable generalization about usage patterns. But A.I. could make that generalization real — tailoring the recommendation and the rationale to each admin's actual account data.

    ✦ Usage-grounded recommendations

    Instead of a blanket Advanced default, an A.I. model could analyze each team's storage, features used, and collaboration patterns — generating a plain-language rationale specific to them: "Your team uses advanced sharing controls and has 2.8TB of active storage. Advanced keeps everything as-is."

    ✦ Dynamic objection handling

    An embedded assistant could surface the questions each admin is most likely to have — based on team size, billing history, and current features — proactively, rather than making them seek out an FAQ.

    ✦ Predictive churn intervention

    A.I. could flag admins at elevated churn risk based on engagement signals — then dynamically tailor the migration message, potentially offering a longer transition period or a contextualised walkthrough without requiring a human escalation path.

    ✦ Natural-language plan comparison

    Instead of a static feature table, a conversational interface could invite admins to ask in their own words: "Will my team lose version history?" or "What happens to shared folders?" — answered from their specific account context, not generic documentation.