Appearance
A is for Access
LettersPractice is a learn-to-read app for families with small children, roughly targeting ages 2+. It's built around daily-ish practice, a minimalist set of exercises, and growing a shared working knowledge between the learner and their caring adult guide(s).
We will have more to say about the product, its design choices, its internals, in future posts. The homepage hopefully carries the best introduction.
This post focuses on our origin, motivations, pricing, and where we imagine ourselves in the existing market.
TLDR: home-grown; $20 one-time payment, with a generous free tier that doesn't expire; effective early literacy at scale; and a good/easy decision for literacy-serious families who are budget or screen-time conscious.
Our origins
Somewhere between 15 and 20 years ago, I encountered the open-source spaced repetition system (SRS) Anki. Briefly, spaced repetition is a learning protocol for fact retention and recall, backed by a strong scientific evidence base. My encounter was well timed - I was taking some introductory Spanish classes, and I used the software to upload the course contents into my brain with extreme ease and efficiency. It was so effective that it left me confused and a little angry that I had gone through 15+ years of schooling without having been taught about it.
Anki was (and remains, I think) the gold standard for power-users, compulsive autodidacts, and medical students. But it is software that was never positioned for mass adoption - instead, it is a power tool that comes with its own learning curve, and asks quite a bit from its users. As my professional life developed, alternating between software development and K-12 education, I remained interested in the wide gap between how effective SRS was for learning, and how niche it remained.
So: around 10 or 12 years ago I began tinkering with my own toy implementation of the same ideas. This was hobby work, and experimentation against a handful of specific ideas for how the SRS paradigm could be made both more user-friendly and more applicable to hierarchical knowledge and skill domains. The core of LettersPractice can be found at this Open Source repository[1].
By around 2021, there was a natural personal convergence. My career in K-12 had given me pretty extensive experience working with low-literacy kids across a variety of contexts. I had a works-pretty-well homemade learning platform. And - the greatest catalyst - I had a three-year-old.
I took what I knew of literacy acquisition, started plugging it into the platform, and worked through it with my daughter on a casual schedule, working against a three-minute timer each day. The results were very good. She began to do serious independent reading around her fourth birthday, and she hasn't stopped since.
The trigger
That might have been the end of it. But some time later, I came across a piece of early literacy software that sells subscriptions at $500 a month. I'm sure it's good, and I want to emphasize that I do not begrudge them, but the price point nagged at me.
Education creates pathways to success for kids of every socioeconomic class. Literacy skills are a powerful lever that enable learning in the practical, literal sense, but also they are also deeply impactful to self-image and identification as "smart".
Here at home we were holding a piece of software that works, and works well, toward the same end. Initial napkin math told me that expected server costs for a given learner's lifetime usage of our app would be $10 or less.
At a bare minimum, I was certain that I could achieve at least half of their results, at about 1% of their price. This alone could greatly expand the early literacy pie, and seemed worth doing.
After a family meeting or two, it was decided. We would take a crack at developing a consumer product, aligned with our own values: evidence-based instruction; low screen time by focusing the app on the pieces of literacy where computers really help; family first, by teaching the family unit rather than the child in isolation; no recurring (forgotten) subscriptions, no ads, and no tracking.
The plunge: enter LetterPeople
The app as it stood was extremely bare-bones. It contained a lot of rough edges that could break its usage, was running with low-quality voice recordings, and required constant adult supervision to keep the user interface from getting into strange states. All of this needed to be upgraded from home-use to product-grade.
But the biggest single gap was in terms of presentation and personality. Real learning - as opposed to passive edutainment - is taxing. For personal in-home use, we did well enough to carry it with parental prodding. But as a product, it felt too dry.
So we've wrapped the curriculum in some personified letter characters. They are the content, but they also present the content. They can be well-behaved and predictable (like v), or a little bit sneaky (like e - so often silent). Like kids, they sometimes find themselves in a friend group where their behaviour goes totally off the rails (looking at you, ough).
Despite inconsistencies in their phonetic work, the letters are unwavering cheerleaders for the kids. They want to be understood, and they speak up at run-time whenever they're able to clarify any observed mistakes.
Our offer
Learning is not a singular event, but a process over time. An app can demonstrate visual polish, and to a certain extent its design sensibilities, over the course of a one-or-two-week trial. But it can't demonstrate the thing that matters most: real learning efficacy.
So our offering is: for all kids, the first 1,000 words are free, and free forever[2]. After that, a one-time $20 unlocks the rest.
What does 1,000 words mean, exactly?
We teach phonetically here. When we say 1,000 words, we mean that we will teach individual phonetic skills that permit the reader to decode at least 1,000 well-behaved English words. In practice, this will mean learning about two-thirds of the simplest letter-sound correspondences, and also a few digraphs (letter-teams like sh, ch, etc).
After crossing the 1,000 word threshold, new facts will be gated, but concepts already introduced will continue to be reinforced and strengthened.
... and our ambition!
We think that our fundamental approach to literacy is good. We have put considerable effort into finding routes optimized against real-world usage, especially through the earliest phonetic ideas.
That said, we don't think that our curriculum sequencing is perfect. We know that we have gaps in our recorded audio base, and we're sure to find lots of surprising quirks along the way.
But we also feel pretty confident that the platform will respond well to experience. It is built to allow pedagogical experimentation and monitoring, to surface relevant aggregate information about observed learning bottlenecks, and to converge over time toward curriculum arrangements that better serve the system's high level goals.
The most effective pedagogy at scale will be data-driven. And data-driven discovery requires wide usage. So to that end, our affordable product is a little self-serving. If we succeed, our curriculum will refine to become still more effective, and our price can fall.
Be in touch!
Thank you for reading.
Try the app. Sign up below for future posts. Or, please, be in touch directly at colin@letterspractice.com.
This repo is actually the third iteration of the software. I had started with a .NET Razor version, then scrapped it for a version built with React, because that was the way the winds were blowing. I eventually decided to ditch the React implementation during Meta's phase of licensing overreach with React. Ah, history. ↩︎
Subject to viability / gross misunderstandings of server scaling costs / etc. This is day one here - I am reserving the right to notice after the fact that I failed to carry a one somewhere! ↩︎