A comment said Google Play is asking them for 12 testers for 14 days. Google wrote that rule for personal developer accounts made after Nov 13, 2023. Mine is new, but it's an organization account, so it didn't apply. Android devs, where did you find your 12?
Five of my bugs in production, ranked by how long they lived: two prices (same day), a win-back email (85 days), upvotes (117), referral codes (207) and the unsubscribe link in my account emails (213). Two had passing tests. Neither test did what a user does.
Same 17 videos on four apps, counted Sep 28, and each app picked a different winner. The one about not charging monthly had 20,800 plays on one app and 99 views on another, where a video about testing AI led with 636.
Squarespace's Core plan went from $23 to $29 a month, billed yearly, in July. You can cancel the site plan and keep the domain, then move it in three steps. Most places won't move one you bought or moved in the last 60 days.
Testers, what window sizes do your end-to-end tests run at? One of mine is now 1272 by 668, a 1280 by 800 laptop once the browser takes its share. It's the window a customer had when the only button that finishes a lesson sat 18 pixels below the edge.
Apple does manually review apps. Mine went through on the third try. The second try came back under Guideline 3.0.0 as one question: did I mean $149.99 and $299.99 a year? A reply would have done it. I sent a whole new build instead and started the review over.
I counted how I talk on camera. In 14 raw takes, 1 in 3 of my lines started with so, and, but or yeah, and I stopped over 2.5 seconds about 11 times a take. Takes averaged 154 seconds for videos that end up about 60. Now every bullet starts with words I say as written.
A makeup artist's site I built runs almost 1,500 tests on every push. My website testing checklist, top five: screen widths, 44 px tap targets, contrast, breaking the form, no dead links. Writing that suite showed every Inquire button went to a page that didn't exist.
The cancel screen on my app's website offered 30% off to people still on a free trial. 0 of 8 took it, which makes sense, since they weren't paying anything yet. Now someone on a trial skips the offer and sees Keep my trial, with the days left.
I set a 50 second default API timeout for AI calls in my app. One feature takes 72 to 102 seconds, so it failed every run for 15 days. Fixing it, I found my retry check never retried a dropped connection, and its test passed because it set the error's name by hand.
On Sep 29 I pulled every comment I could from 30 days of posts into one list: 73 from other people. 34 are on one video, the one about not charging monthly. 3 read like a possible customer, so the list pulls those out. I still type every reply.
The .com domain price increase starts November 1. The registry fee goes from $10.26 to $10.97, so at cost it's $10.46 now and $11.17 after. You can renew up to 10 years out at today's price. While you're in there, check whose name it's in.
My website's contact form has one optional question: where did you find me. I want to know which videos bring in work, so skipping it costs nothing and the form still sends. Small business owners, where did your last five customers find you?
My best video has over 20,000 views, and my reply to its comments has 237. The reply ran 3 minutes 26 seconds and answered all 46 comments at once. Now every video is under a minute with one idea, and I'll keep posting the real numbers.
My app's website speed dashboard said time to first byte was 2.67 seconds that week. Nothing I shipped slowed it. A third of the samples came from one five-hour burst, and four login page loads took 76 to 299 seconds to paint, from background tabs.
I wrote a program that edits my talking head videos, and it knows when I'm reading my notes. A gaze model reads a glance at them as 20 to 55 degrees and eyes on the lens as under 6, so every cut starts when I'm looking at you. Can you catch me reading in this one?
Three ways to pay for a website from me: half now and half when it's live, over four months with no interest, or all at once for $50 to $100 less. The October price, 30% off, is the first way only, with half down by Oct 31.
A customer told me a button didn't exist. His window was 668 pixels tall and Mark done started at 686, so he spent 50 minutes on the page and never clicked it. Now the button sticks to the bottom, and a test opens the page at his exact size.
On Sep 28 I had 153 posts queued across seven apps, 125 video posts and 28 text posts, through Nov 1. The next day I wrote a check for the calendar. It found a text post 3 hours from a video (my rule is 4), and a draft that put two videos a day on two apps for 12 days.
My bulk deal cost more than paying by the hour. Changes after launch are $70 an hour, and I was selling four hours for $315, when four at $70 is $280. The block is $240 now and never expires, and a test fails if the math ever breaks again.