Designing for user quality, not user volume
Role
I owned the initiative end-to-end as a senior product designer: pitched it, led the research, designed the solution and analyzed the post-rollout data. One PM collaborated on development and scope, developers helped to optimize design for development, marketer reviewed copy.
Problem
Users were completing sign-up but dropping off at activation. The obvious UX instinct was to reduce friction, but the data pointed somewhere else.
Solution
Two parallel changes: moving friction from activation to the sign-up funnel (where it filters out non-viable users earlier), and upselling hesitant ones from free account to free trial so they convert better downstream.
Outcome
Time to value dropped by roughly 40%. End-to-end activation rate grew from 16.86% to 20.62%, a 22% relative lift. Trial-to-paid conversion grew by about 14% relative. Engagement improved across the funnel (publishes +3pp, embed copies +13pp, videos reaching 10+ plays +15pp).
Context
Vidalytics is a profitable B2B video marketing platform with around 50 employees and roughly 2,200 paying customers across monthly subscriptions and enterprise contracts. Our core audience is VSL, sales, and marketing professionals whose income depends on video performance, so they tend to arrive highly motivated.
I had been at Vidalytics for about two years as the senior product designer on a small product team (two designers and two PMs). This initiative started with routine competitor research which led to full-scale initiative which changed how we perceive our sign up and activation funnel and increased ARR by ~5-6%.
The strategic bet
The core call here was unusual: we deliberately accepted lower sign-up conversion in exchange for higher-quality activations and better paid conversion downstream.
Most funnel optimization defaults to "reduce drop-off at every step." Our research showed that wasn’t the right move for us. A meaningful portion of our drop-offs were non-target users leaving on their own, and that was actually working in our favor. However, the friction that filtered them out sat at activation, which is the same place where viable users were struggling and taking a long time to activate or sometimes even dropping off. The real opportunity was to move that friction upstream into the sign-up funnel, where it would filter non-viable users earlier while reducing the barriers that made viable ones struggle at activation. On top of that, we could use the sign-up flow to nudge hesitant-but-viable users from free account to free trial, where they’d get more value and convert at much higher rates.
That idea: trade volume for quality, is the one I had to pitch, defend, and prove to the rest of the team.
Discovery
Our legacy sign up and activation flow. Unfortunately, there was no video recording of this flow in real application, so i had to remake it in figma prototypes.
While auditing how profiling was performing, I analysed our sign-up and activation funnel in Mixpanel. Drop-off rates at activation steps (upload initiation, embed copy, publish) sat above industry benchmarks, and well above the funnel’s other steps:
That combination was unusual. Users who complete sign-up typically engage with core features at least once. The standard explanation would have been UX friction, however a competitor scan of Vimeo, Wistia, VideoAsk, Vturb, Brightcove, Vidyard, JW Player, and Kaltura didn’t support it. Our sign-up flow wasn’t meaningfully harder than peers targeting similar audiences. On the contrary, some competitors that are more enterprise-focused presented even more frictionful funnels (either requiring a personal demo, sign-up approval, or only offering paid account options).
Since the UX hypothesis didn’t fit, I pitched an investigation to the senior PM in Slack, attached the Mixpanel numbers, and got a green light to spend research time on it after discussing this more thoroughly on weekly product department meeting.
Example of competitor sign up and activation flow: Wistia.
Research: testing every plausible explanation
I built five hypotheses for why motivated users would abandon after sign-up:
Technical Issues
Friction Exceeding Perceived Value
Confusion
Mismatched Expectations
Lack of Trust or Commitment
Each needed ruling in or out with real evidence.
Ruling out technical issues
I partnered with QA to audit error logs, which did not present any issues which might cause this pattern to arise. Then I searched six months of Intercom tickets across 56 activation-related keywords, and only 5 tickets were relevant (all already-resolved third-party issues with oAuth or payment). Which is expected as users who do not activate rarely reach out to support, but to be confident i checked it out as well. Finally, I watched 100 session recordings of users who dropped off at activation, which is about 25% of new accounts in the analysis window.
Some issues appeared in 12 of those sessions, and 9 of them were self-resolved in the same session. Most issues were related to input format. That's not a plausible cause of abandonment.
As a side finding, about 12% of users encountered some form of technical friction during sign-up, most of it self-resolved. Not relevant to this initiative, but worth flagging for a separate investigation later, particularly around how we display input fields.
Ruling out confusion
Funnel timing showed the opposite: drop-off users completed steps anywhere from ~30% to ~75% faster than users who converted, averaging ~40% faster across all activation steps. Mixpanel's Flows tool showed less than 3% of them navigated back and forth. Instead, users hit activation steps and left cleanly, with no further app interaction. A second batch of 100 session recordings confirmed the same pattern.
So users who dropped off weren’t struggling. They were leaving fast, and they weren’t coming back.
The turning point: ruling in user-market fit
Drop-off users spent less time on each step, not more.
Session recordings showed no friction, no confusion, and clean exits.
Channel segmentation was decisive. Users from player links converted at 0.02% through the full funnel, referral at 0.069%, compared to 1.55% for organic search. Important to note that these are the numbers of funnel starting from the website landing page. This is the only place where funnel start this early because our Mixpanel only tracks marketing channel correctly with this step included. As confirmed by previous data analysis, first two groups are channels through which less motivated users come and organic search is the channel with more motivated audience.
Pre-funnel behavior analysed using mixpanel flows tool sealed it: low-converting users spent almost no time on the website, didn’t explore pricing, and didn’t browse other pages before clicking sign-up.
Two levers, two interventions
The research pointed to two separate opportunities:
Move friction upstream: the filtering was already happening at activation, but that’s exactly where viable users were struggling too. By shifting that friction into the sign-up funnel (through an upsell and reduced free-account visibility), non-viable users would filter out earlier, while viable users would face less friction at the activation steps where they were losing momentum.
Upsell to free trial: since free trial users converted at about 25.43% versus 3.2% for free account, nudging hesitant-but-viable users toward trial instead of free account would give them more value and increase their likelihood of converting to paid.
These two levers worked together. Moving friction upstream would both filter non-viable users earlier and reduce friction at activation for viable ones. The upsell to free trial would make sure that the viable users who did get through would be on a plan where they actually get full value, which converts much better.
The trial-to-paid conversion gap (around 25.43% vs. 3.2% for free account) was the data point that made the upsell angle clear: users on free trial got more value, and that translated directly into conversion.
Solution 1: strengthening the sign-up filter
The sign-up flow needed two small, low-risk changes to filter harder without pushing real prospects away.
Reduced visibility of the free account option. It's still available but is shown below the fold of the screen: before vs after of our pricing page.
A personalized upsell based on profiling answers, nudging users toward free trial instead of free account. Title, body text and bullet points in this upsell are changing depending on which answers user picked during profiling. In total there are 7 different title + body text combinations and 16 different bullet points.
Both changes were intentionally conservative. If the hypothesis was wrong, they wouldn’t cause meaningful drop-offs among real prospects. If it was right, they’d prove the thesis and open the door to bolder bets later (for instance, testing the removal of free account entirely).
Solution 2: Upload First Video, redesigning the activation path
The bigger design intervention sat after sign-up. Viable users were still hitting friction at upload and embed, which are the critical activation moments. The existing experience was fragmented: sign-up, profiling, a generic welcome page, a walkthrough video most users skipped, an overlapping Appcues tour, and only then, navigation to upload. By the time users reached the upload step, many had already lost momentum.
Competitor analysis reinforced this. Wistia routes users directly to upload with a brief, plan-tailored video inline. VideoAsk uses an empty state with a clear Create CTA. Vimeo skips the welcome experience entirely on the free tier. The shared pattern across all of them: one focused next step, not a stack of overlapping onboarding layers.
So I rebuilt the post-sign-up flow around one question: how fast can we get a viable user through upload and into vid settings, where the first real aha moment happens?
Our new sign up and activation flow. Pricing page redesign to hide free account option, personalized upsell introduced and upload path made more frictionless.
Welcome modal with a promo video under one minute
Marketing wanted a promo video in the flow to show what a Vidalytics video looks and behaves like in practice. The previous version of the welcome video was mostly ignored, so we needed a way to incorporate this one without repeating that pattern. I picked modal over a dedicated page because it doesn’t interrupt the primary flow and is overall a more frictionless approach as approved by public research. The user lands on upload, which is the prerequisite step that unlocks the first real value moment (vid settings), and the welcome surfaces over it. For video duration, marketing team decided short 1-minute format would be best based on our video-performance database (Vidalytics runs on video analytics, so we have a deep well of data on how users engage with short-form marketing video).
Upload page as the landing destination
After the welcome, users land directly on upload, not dashboard, and not a generic home. I considered dashboard and rejected it, because a dashboard with no uploaded videos is an empty state with nothing useful to show. Upload is the fastest path to the first aha moment: once a video is uploaded, users land on vid settings, where they see the full range of customization options, interactive elements, and marketing adjustments that make Vidalytics valuable. Upload itself isn’t the value, it’s what unlocks it.
After going through the full sign up flow, user gets straight to the upload page. Before users landed on welcome page with video and CTA to upload a video which introduced friction with no additional value.
Lower-friction upload paths
I added multiple upload sources (integrations, cloud drives, direct URL) so users could start from wherever their video lived, rather than forcing a single-source flow.
Upload page before and after.
Demo video fallback
For users without a video ready, I added a pre-configured demo so they could still experience the core settings and interactions. The intent was to let users explore the product on real content instead of getting stuck at a hard requirement. This would also be a good example of our settings capabilities presenting all the range of adjustments and features we have.
A demo video is presented to users so they could see what Vidalytics settings and analytics are capable of before gathering the real data and figuring out how settings work which usually comes a bit later in the user lifecycle.
Smart routing for returning users
Non-activated returning users land on upload (so they can pick up where they were). Activated users land on dashboard (their actual workspace). The surface meets users where they are in their lifecycle.
What I chose not to do
Remove free account entirely. I considered it but decided to test the smaller hypothesis first. Removing a free tier has long tails, and confirming the directional thesis with low-risk changes felt like the more responsible first move.
Build a fully native onboarding from scratch. Same reasoning. Prove the strategic bet before investing in a larger system.
Optimize specifically for mobile. Mobile is 7 to 9% of our sign-ups, and our internal flow analysis showed mobile sessions concentrate around specific check-in behaviors (stats review, video status) rather than the core flows (video setup and proper analysis). I made sure the new flow functioned correctly on mobile, but I didn’t spend research and design cycles on a cohort whose usage pattern didn’t match the core journey. (This is now being addressed as a separate initiative which another designer will be leading, which I cover in What’s next.)
Run formal external user testing. At our company scale, and given the solution's low engineering complexity, proper user testing would have cost more than the potential launch cost and risk. We ran internal team usability review, iterated twice on the designs, and after PM sign-off pushed it to production.
Selling the strategic bet internally
The thesis was unusual enough that it needed more than a Slack thread to land. I ran a follow-up presentation to the product department at all-hands (both PMs and the other designer), walking through the research, segmentation, and the quality-over-volume argument.
The sharpest pushback came from the other PM. The concern was this: if we reduce visibility of the free account option, we’ll lose users who would have eventually converted via free account, and we’ll gain less money from pushing hesitant users to trial than we’ll lose from reduced free-account sign-ups.
It was a fair challenge. I addressed it directly with the conversion gap in the data: free account users converted to paid at about 3.2%, while free trial users converted at about 25.43%, roughly a 7x difference. Even with significant friction from the upsell, the math favored the trial path by a wide margin. The low-intent users we’d filter out weren’t meaningfully contributing to revenue anyway.
The senior PM approved the initiative at the close of the meeting.
Results
For the performance analysis, I used Mixpanel’s funnel tool to compare pre-rollout and post-rollout cohorts data, filtering out internal users for data clarity. I built activation funnels for each cohort (before and after the feature launch), then segmented by device, plan type, and marketing channel to see if the changes affected different groups differently. And here are the results:
A note on cross-impacting initiatives
Two other features shipped during the measurement window that I need to call out before jumping to results that we achieved:
Sidebar Upgrade Promo and Trial/Upgrade Flow (launched the same day as this feature with no prior a/b testing), designed by our other designer, but part of the broader activation and conversion strategy our team was shaping at the time. This had a noticeable effect on conversion to paid, so some portion of the trial-to-paid and free to trial lift can’t be cleanly attributed to the sign-up/activation changes alone.
Home Stats (launched 2 weeks after rollout), which I designed. Primarily affects retention on sign-in, vid stats enter and vid settings edit, with a marginal effect on conversion.
I isolated the period before Home Stats was launched and the numbers came out similar to the whole analysed period, with differences small enough to consider negligible. So we can assume most impact to analysed metrics came from the features of this initiative. When it comes to the sidebar promo, both sidebar and activation optimization initiatives were part of the same quarterly activation OKR, and the honest attribution is hard to tell since their launch dates match. The sign-up/activation redesign and the upgrade promo each contribute meaningfully to the observed lift, and the exact split isn't cleanly separable without a more complicated analysis, which we didn't have the data-team capacity to run.
Time to value and conversion quality
Median time to activate dropped from 20.1 hours to 12.1 hours, roughly a 40% reduction. This is one of the clearest signs that the flow redesign worked as intended. By moving friction upstream into the sign-up funnel and streamlining the path from sign-up to upload to vid settings, viable users reached the first aha moment (vid settings) much faster. Less time spent navigating between disconnected onboarding layers means less momentum lost along the way.
Trial-to-paid conversion grew from 25.43% to 29.05%, about a 14% relative lift. This growth comes from two reinforcing effects. First, the upsell shifted more users from free account to free trial (the ratio moved from roughly 58/42 to 62/38 in favor of trial), so a larger share of new users were on the plan type that converts better. Second, the faster activation path meant trial users experienced the product's full value sooner, before their trial window ran out, giving them a stronger reason to convert.
Overall activation rate moved from 16.86% to 20.62%, a 22% relative lift. The downstream improvements compounded: even though we deliberately added friction at sign-up (upsell, reduced free account visibility) and saw fewer users start the upload step, the users who did reach upload were more motivated and went substantially further. The denominator shifted to a more motivated cohort, and that cohort converted at higher rates on every step that follows upload.
Activation funnel, step-by-step
Total funnel conversion: 16.86% vs 20.62% with such breakdown by steps:
The pattern: fewer users started, but those who did went much further and got more value out of the product. The drop at upload initiation (62% to 47%) combined with gains on every step after it (publish, embed, 10+ views all up) is consistent with what we expected: the sign-up flow changes filtered out lower-intent users before they reached activation, which means the users who did reach upload were more motivated and followed through at higher rates. End-to-end activation rose from 16.86% to 20.62% as a result.
Paid vs free cohort breakdown
The most interesting finding in the cohort breakdown is the asymmetry between free trial and free account users. Free trial end-to-end activation went from roughly 24% to 30%, about a 1.25x lift, while free account activation barely moved (from roughly 11.56% to 12.37%, about 1.07x). Both technically improved, but trial users responded meaningfully to the changes while free account users essentially didn't.
That asymmetry tells us something important about the intervention. Trial users signed up with enough intent to commit to a time-limited evaluation, and when we reduced friction at activation (upload-first flow, fewer overlapping onboarding steps, faster path to vid settings), that intent translated into action. Free account users have less urgency by design, and even with a smoother path they didn't move through it more decisively. Friction reduction by itself isn't enough to convert low-urgency users, the path has to meet motivation that already exists.
This validates the upsell decision more strongly than expected. By nudging hesitant-but-viable users from free account toward trial, we weren't just shifting plan labels, we were moving them onto the path where the flow improvements had nearly all of their effect. Without that upsell, the activation redesign alone would have done much less for free account users than it did. That tells us the intervention didn't change where users struggle within the funnel, it changed who reaches the funnel in the first place.
Desktop vs mobile
Desktop conversion improved from 19.58% to 24.86%. Mobile barely moved (roughly 5.8% to 5.0%), and the mobile sample is too small to draw conclusions from anyway (40 and 35 users entering the funnel respectively, with fewer than 5 converting in each period).
On mobile, most drop-offs concentrated on the upload initiation step more than on desktop, with other steps following the same distribution as the general funnel. Given that mobile is 7 to 9% of sign-ups and the primary mobile use case is check-in behavior rather than full creation flows, this is something to address separately rather than a signal about this initiative’s performance.
Business impact
Combined, the higher activation quality, faster time to value, and stronger trial-to-paid conversion translate to an estimated 5 to 6% ARR lift for the business. The calculation works like this: the trial-to-paid conversion lift (25.43% to 29.05%) applied across our monthly sign-up volume produces a measurable increase in new paying customers per month. That increase, multiplied by the average revenue per customer over a year, gives the ARR contribution. On top of the direct conversion lift, the upsell shifted the free-to-trial ratio from roughly 58/42 to 62/38, which means a larger share of new users are now on the plan type that converts at roughly 7x the rate of free accounts, compounding the effect over time.
It is worth noting that this estimate combines gains from the sign-up/activation redesign with contributions from the parallel Sidebar Upgrade Promo shipped in the same window (see confounding factors above). The honest attribution between the two is hard to separate cleanly, so the 5 to 6% figure represents the combined impact of the broader activation OKR, not this initiative in isolation.
What didn’t work
Two of my design decisions produced disappointing data. I’m keeping them here because the reflection matters more than the polish.
The welcome video
Welcome video modal which is shown when user lands in the app.
I introduced the welcome modal with a promo video. Post-rollout, only 12.97% of users who saw the modal started the video (48 out of 370 accounts). Of those who started, 83% dropped off by second 7, and just 4% watched to the end. Average watch time was 8.2% of the total duration. The unmute rate was 50.57%, and the bounce rate (users who started but closed within the first 5% of the video) was 83.55%.
Retention data supports that read. Welcome video watchers retained at nearly the same rates as the general average across all tracked actions: sign-in (19.75% vs 19.76% avg), vid settings change (20.38% vs 20.88% avg), and publish (17.44% vs 20.66% avg). If the video was actually changing user behavior, we would expect to see a retention difference. We don’t. That tells us the welcome video isn’t hurting anything, but it isn’t helping either.
Overall it seems that the welcome video needs either significant rework, personalization by segment, or removal.
The demo video fallback
I added a demo-video path for users without a video ready, reasoning they could experience the product on pre-configured content and see the vid settings capabilities without needing their own video. About 7.1% of users chose this path instead of uploading. Free and trial users chose the demo option at similar rates (16 to 18%), so the choice wasn’t plan-type specific.
The demo path underperformed on every metric we track. End-to-end activation was 1.65% for demo users vs 20.62% for the upload path. Conversion to paid was 22.22% vs 29.05%. Retention was 2 to 2.5x worse across tracked actions.
The step-by-step funnel comparison shows where the gap opens up. For trial users on the upload path, 80.66% complete the upload, 90.06% publish, 58.44% copy embed, and 95.56% reach 10+ plays. For trial users on the demo path, only 46.34% even complete the demo flow, then 78.95% publish, 40% copy embed, and 83.33% reach 10+ plays. The pattern is consistent for free users but more extreme: only 31.82% of free demo users complete the flow vs 89.55% on the upload path. On mobile, demo users essentially don’t complete the funnel at all (zero completions past publish from a sample of 3 users who started).
The biggest drop happens at the very first step: completing the demo flow itself. That tells us the problem isn’t just about what happens after the demo, it’s that users struggle to get through it in the first place. Two plausible explanations: first, users who don’t have a video ready may not be active video marketers, which means choosing “try the demo” may itself be a signal of lower intent rather than a neutral preference. Second, the demo flow may have usability issues I didn’t catch in internal team testing, where testers already understood the product and could navigate it regardless of the flow. In either case, my original reasoning (let users explore before committing their own content) doesn’t hold up in the data. The demo path underperforms on every metric, and it’s a candidate for rework or removal in a future iteration.
What this initiative taught me
The most important finding from the research was unexpected: the biggest drop-off pattern in our funnel wasn’t a UX problem, it was a user-market fit problem. Standard design instinct is to reduce friction at every step. That instinct, applied unquestioningly, would have made the funnel worse, because the friction at activation was doing double duty: filtering out non-target users while also slowing down viable ones. The answer wasn’t to remove friction, it was to move it to a place where it only filtered, without blocking the users who should be getting through.
That reshapes how I approach funnel work now. Before asking "how do we fix the drop-off?", I ask "do we want to fix this drop-off, or is it already doing something useful?" Not every number that moves in the wrong direction needs correction.
The second lesson is calibration. Paid-path users responded about 2x more strongly than free-path users to the same activation changes. One reading of this is that motivation compounds with UX improvements, where small friction reductions help motivated users more. Other readings are plausible too (sunk-cost commitment, different behavioral baselines by cohort, different upstream friction). Whichever interpretation holds, the takeaway is the same: design interventions and target cohorts have to be paired, not treated as independent variables.
The third lesson is that not every design decision lands, and owning that is part of the job. The welcome video and demo video both underperformed. This case study is stronger for showing that honestly rather than selectively.
What’s next
This initiative directly informed several follow-ups:
Mobile Optimization: Currently in research phase, which I'm leading. Once the use case is validated, design will be handed off to another designer to implement. Mobile accounts for 7 to 9% of sign-ups, and converts much worse than desktop.
Onboarding Optimization: A broader follow-up initiative already underway, directly downstream of this work.
Welcome Video Rework Or Removal: Candidate for the next iteration, based on the data above.
Demo Video Rework Or Removal: Same reasoning as with welcome video.
Pushing The Filter Further: Now that the thesis is confirmed, the original "remove free account entirely" alternative is back on the table as a future test.