Connie Liu
Product Manager for mobile at Rev, interviewed by Steve P. Young of App Masters
Where to hear and read about Connie Liu1 episode or test
Connie Liu is the product manager for all of Rev's mobile products, which serve over 100,000 monthly active users. The thing most worth learning from her is to treat every step from store impression to paying customer as its own conversion rate, and to find what to fix by watching confused users.
Usually confusion is is the number one drop-off reason.
How App Store Screenshots & Preview Videos Increase Downloads
Connie Liu: the story and the lessons
1Connie Liu and the scale of Rev's mobile apps
Liu manages all of Rev's mobile products. Together the apps have over 100,000 monthly active users and bring in millions of dollars in revenue.
At that scale, small gains in conversion matter, which is why she works on store page and in-app funnel details.
Liu fixes store conversion before buying more traffic. Traffic that doesn't turn into installs is "a waste of money," so anyone who lands on the app page should be likely to install.
Her order of work is to make the page convert first, then scale acquisition.
2How Connie Liu rebuilt Rev's App Store page
Liu redesigned screenshots with big text and vivid colors. The old screenshots looked dated and had small fonts that didn't stand out in the scrolling preview. The new ones let people read the value prop without opening the page.
Rev's impression-to-install rate rose over 82% after the changes it tried, including the new screenshots.
Liu put social proof in the third screenshot panel. It shows without scrolling and highlights name-brand customers and the total customer count. User tests found people trust the app with recognizable users when apps look alike.
The tests compared apps with similar features and similar review counts. Steve Young's client reported a 91% download gain from better screenshots and social proof alone, with no keyword changes.
Liu pushed up the number of ratings, not just the stars. User tests showed people look at the count of ratings as a trust signal. Rev added an in-app trigger asking users to rate, and one app went from hundreds of ratings to many more.
She puts it with an analogy: "would you rather go to a restaurant that looks empty or a restaurant that's jam-packed?"
Liu added an autoplay preview video to one Rev app. The video plays by itself in search results, while static screenshots don't. Rev compared the app page with the video against the one without and saw a lift.
The video went on a different Rev app from the one whose screenshots were redesigned.
3How Connie Liu turns installs into paying customers
Liu treats each funnel step as its own conversion rate. Rev tracks install to recording, then recording to order. Recording means making an audio recording, the first action Rev wants.
The end goal is a paying customer, not just an active user. Her first rule at every step is to learn from users first.
Liu watches real users to find where confusion causes drop-off. Rev runs user tests and watches people use the app, asking where they get confused. Users who don't know what to do will leave.
What she sees becomes a hypothesis for a launch or an A/B test.
Liu reads support chats for missing information. Rev has Intercom in every app and reads the questions users send. Users kept asking "what is this transcription that you keep talking about?"
That told Rev the app didn't explain transcription well enough, and it became something to test.
Liu cut the empty recorder screen down to one action. The old screen offered a sample recording, trim, record, export, search and many settings. Users didn't know where to focus, and many left before seeing the app's value.
The new screen has a light bulb graphic and one big record button. Once users start recording they feel the value, and Rev can lock them in.
4The checkout priming screen Connie Liu's team removed
Rev's extra screen before checkout tanked conversion. Rev earns money when users pay to transcribe. It added a priming screen with information about transcription right after users tapped transcribe, hoping education would nudge them to buy.
Liu says "we made the mistake of kind of introducing a checkout priming screen." Rev removed it.
Liu weighs friction against added information and tests both. Other Rev experiments showed extra information beating the cost of friction, but here a whole extra screen was too much.
Putting the information on the checkout page itself might have worked. Her caution is that any added friction can tank conversion.
How App Store Screenshots & Preview Videos Increase Downloads
Rev's mobile PM explains how redesigned App Store screenshots, social proof, a preview video and a simplified first screen lifted conversion at each step from impression to paying customer. She also shares a checkout priming screen that failed and tanked conversion.
Liu manages all of Rev's mobile products. Together the apps have over 100,000 monthly active users and bring in millions of dollars in revenue.
At that scale, small gains in conversion matter, which is why she works on store page and in-app funnel details.
Liu fixes store conversion before buying more traffic. Traffic that doesn't turn into installs is "a waste of money," so anyone who lands on the app page should be likely to install.
Her order of work is to make the page convert first, then scale acquisition.
Liu redesigned screenshots with big text and vivid colors. The old screenshots looked dated and had small fonts that didn't stand out in the scrolling preview. The new ones let people read the value prop without opening the page.
Rev's impression-to-install rate rose over 82% after the changes it tried, including the new screenshots.
Liu put social proof in the third screenshot panel. It shows without scrolling and highlights name-brand customers and the total customer count. User tests found people trust the app with recognizable users when apps look alike.
The tests compared apps with similar features and similar review counts. Steve Young's client reported a 91% download gain from better screenshots and social proof alone, with no keyword changes.
Liu pushed up the number of ratings, not just the stars. User tests showed people look at the count of ratings as a trust signal. Rev added an in-app trigger asking users to rate, and one app went from hundreds of ratings to many more.
She puts it with an analogy: "would you rather go to a restaurant that looks empty or a restaurant that's jam-packed?"
Liu added an autoplay preview video to one Rev app. The video plays by itself in search results, while static screenshots don't. Rev compared the app page with the video against the one without and saw a lift.
The video went on a different Rev app from the one whose screenshots were redesigned.
Liu treats each funnel step as its own conversion rate. Rev tracks install to recording, then recording to order. Recording means making an audio recording, the first action Rev wants.
The end goal is a paying customer, not just an active user. Her first rule at every step is to learn from users first.
Liu watches real users to find where confusion causes drop-off. Rev runs user tests and watches people use the app, asking where they get confused. Users who don't know what to do will leave.
What she sees becomes a hypothesis for a launch or an A/B test.
Liu reads support chats for missing information. Rev has Intercom in every app and reads the questions users send. Users kept asking "what is this transcription that you keep talking about?"
That told Rev the app didn't explain transcription well enough, and it became something to test.
Liu cut the empty recorder screen down to one action. The old screen offered a sample recording, trim, record, export, search and many settings. Users didn't know where to focus, and many left before seeing the app's value.
The new screen has a light bulb graphic and one big record button. Once users start recording they feel the value, and Rev can lock them in.
Rev's extra screen before checkout tanked conversion. Rev earns money when users pay to transcribe. It added a priming screen with information about transcription right after users tapped transcribe, hoping education would nudge them to buy.
Liu says "we made the mistake of kind of introducing a checkout priming screen." Rev removed it.
Liu weighs friction against added information and tests both. Other Rev experiments showed extra information beating the cost of friction, but here a whole extra screen was too much.
Putting the information on the checkout page itself might have worked. Her caution is that any added friction can tank conversion.