OpenLiteracy Phase 2 · v1 MVP Student-Lesson Journey
Scope: this doc covers the v1 MVP student-lesson journey only — a student enrolled, taking lessons in sequence. It does not cover catch-up lessons, multi-group flow, or the full educator surface. Those will be modeled separately.
For every diagram, the Current side shows what's in the Phase 1 code today. The Proposed side shows what we're recommending based on what your team described in the 5/04 intro and 5/11 kickoff. Where a proposal is grounded in a direct quote from your team, the source appears below the diagram in a green box so you can see exactly where it came from.
This is the agreed-upon v1 MVP journey for a student going through lessons — not the full picture (no catch-up lessons, no multi-group flow). Everything below already exists in code and is what we all agree on. No questions about this diagram — we are not asking you to review it. Sections 1–7 below are where we need your input.
The v1 MVP end-to-end flow. This is what the Phase 1 product does today. The sections that follow drill into each decision point and show what we're proposing to change.
Score bands and lesson part-structure — these show up everywhere downstream.
The big difference: the proposed system splits the middle band into two — a score of 8 advances with a warmup attached; a 6 or 7 stays in the lesson and remediates. RED becomes "drop down to the previous lesson" rather than just stopping. This 4-band split is our proposal for you to confirm or revise.
Our assumption is writing/dictation and fluency stack get dropped or merged into the other parts. Confirm in N.2.
"We're only going to have four parts to each lesson, not six." Sarah Scott Frank — 5/04 intro meeting, 9:30
The placement screener — once when a student enrolls. Determines what skill they start on.
The proposed change adds a sentence-level check first, so strong readers don't have to grind through 10 words on every skill. The student only reads the word list if they miss a target word in the sentence.
"We want all students to start with this skill, CVC… And what we want to shift from is from reading words to reading a sentence. So every skill will start with reading one sentence, and then based on how they perform on that sentence, which will have target words within it, they would then move into needing to read a word list." Rebecca Patterson — 5/04 intro meeting, 18:06
"It will be 100% for the sentence to be like, okay, that skill's green and we're not going to worry about it, we're going to move on. But if they get even one of the five target words wrong, then they would move to the word list, and the word list is going to be like 90% then becomes green." Rebecca Patterson — 5/04 intro meeting, 19:24
"The sentence piece is so that those more advanced readers are just reading like 10 sentences, and then they're done for the whole assessment, and they're going to get started with that learning at their level." Greta Phillips Kendall — 5/04 intro meeting, 19:48
Today's screener stops on the first skill where a student scores 0–5. The proposed rule keeps the student going through one more skill if they bottom out completely — only stopping after two consecutive zero-zero results. This is our interpretation of what Rebecca described at the 5/04 kickoff (quotes below). The intent we picked up: catch students who genuinely can't read at the level we started them on, while still giving them a chance to demonstrate strength in a higher skill they might unexpectedly know.
"They would go on to do the digraph sentence and so on until they meet that discontinue criteria, which is two in a row… That's zero sentence, zero words." Rebecca Patterson — 5/04 intro meeting, 27:00
"We don't want to stop if they're red, because we could have a lot of reds, but they're still reading some words. We want to stop if they're like zero, zero, like totally bottom-out and out, just for the discontinuity." Rebecca Patterson — 5/04 intro meeting, 28:13
"We were a little crayon initially, but the zero, zero is what we want." Rebecca Patterson — 5/04 intro meeting, 1:15:15
How a single lesson gets composed for different session lengths. Key clarification from OL (5/20): session duration is set when a group is created or in group settings, not chosen at lesson start. Once a group has a duration, every lesson for that group uses the matching template.
Duration is a property of the group, not the lesson. Teacher picks it at intake (when assembling the group), can edit it later in group settings. Lessons then load the matching template automatically.
"We talked about, part of the scope is like, you could adjust the time. So you could have a 15-minute lesson. What is in a 15-minute lesson? Is it that we're splitting a 30-minute lesson in half and it becomes two days?" Sarah Scott Frank — 5/04 intro meeting, 9:30
"System recommended will not work. Schools have schedules. Recommended that the duration is set as part of intake — when they are specifying the schedule, set the duration of a lesson as part of intake." Your team's response to our Round 2 questions on session duration
Warm-ups and lesson skip logic. We think of these at the group level, not individual student — a group can be as small as 1.
What is a "per-pattern matrix"? A grid where the rows are skills and the columns are word-patterns within each skill. Each cell tracks whether the group has demonstrated mastery on that specific pattern (green) or not yet (red). The warmup pulls from the red cells.
"You could serve up a warm-up that's really targeted towards actually one kid in the group, and you could serve up two practice activities that's targeted towards the vast majority of them and this other outlier… you could somehow notate for the teacher, 'this practice makes sure that this is really served up to John'… this group has mixed needs." Sarah Scott Frank — 5/11 Sprint 1 Kickoff, 25:30
"You could imagine building a different database that was more of a matrix… you just had both a horizontal and a vertical that you were going from." Sarah Scott Frank — 5/11 Sprint 1 Kickoff, 8:43
Implementation note: the per-pattern matrix is new structure. It needs storage (matrix per group), a definition of how cells move to "mastered," and a word-pattern source list. We have a question open on whether your team can supply the word-pattern source list (C.3).
Today's product runs lessons in strict sequence — a student takes lesson 28 even if they already demonstrated mastery of W in the screener. The proposed change uses the per-pattern matrix to skip lessons whose target is already mastered.
"If they know 10 letters, can we skip some of those letter lessons because they know J and W and T and M and S and P? And so maybe we can focus on it… letter level and at CVC, like if you know short E and short A, but you don't know short I and short O, then we can really focus our attention on I and short I and short O." Sarah Scott Frank — 5/04 intro meeting, 38:27
"If you think of this instructional time as gold, you have this amazing opportunity to move kids really fast. If they already know the skill, the worst use of a kid's time is when they're just marching through lessons that are too easy." Sarah Scott Frank — 5/04 intro meeting, 40:01
When is a group ready to move from one skill to the next? What triggers a formal "progress monitor" check?
Today's progression is "next lesson_number = current + 1." No distinct "skill transition" event, no progress monitor as a separate concept.
E.8 asks you to confirm the trigger conditions and adjust the thresholds. The triggers below are drawn from what your team described in the 5/04 intro.
"Ways that it can be triggered. So it would be like a different outside of the lesson. We wanted like a button that's always available, progress monitor. So then that's educator choice. I want to see. I have a meeting coming up. I think they're ready. Whatever the use case is. Or that 90%, you've excelled through three lessons in the skill bucket. You don't have a choice. Let's see if they're ready." Rebecca Patterson — 5/04 intro meeting, 56:59
"I think we said 80% would move forward." Rebecca Patterson — 5/04 intro meeting, 56:46
"The progress monitor is going to be reading 10 words to kind of exit the skill. So it's like the exit ticket, but it's going to be a new list… five real, five nonsense that encompasses all of the skills in that skill bucket. The progress monitor hasn't been built yet." Rebecca Patterson — 5/04 intro meeting, 55:13
B.8 — without a "turn Green" rule, Yellow skills accumulate and never resolve. Our recommendation is Option A (next progress monitor clears it), which matches what your team described at kickoff.
"Think about those skills in the initial assessment and some were yellow, some were red — the progress monitor is going to be the only thing that now makes it green. Now that's green. And the idea is to obviously by the end of the year have as many green as possible. So it takes a skill they didn't previously pass, makes it green, shows that they're ready to move on." Rebecca Patterson — 5/04 intro meeting, 57:31
Your team identified remediation as the area that most needs redesign. This section walks through the current behavior and what we're proposing.
"I think, on the back end, our logic isn't totally right. Our remediation is weak." Sarah Scott Frank — 5/04 intro meeting, 9:30
"The hard part is the remediation because, what is the AI tool going to generate to remediate for the students?" Rebecca Patterson — 5/04 intro meeting, 12:20
Two gaps: (1) no upper limit on attempts — group can theoretically be stuck forever, (2) every retry generates a brand-new AI lesson; no "retry same lesson with new practice" pathway.
Two changes vs. today: (1) tier the remediation by how badly the group missed; (2) cap the number of attempts before alerting a specialist; (3) on the first retry, reshuffle the same lesson (different practice words, same teaching) rather than generate a fresh AI lesson — only generate new on attempt 2. Please adjust the tier thresholds and the cap.
D.7 — repeated exposure to the same passage is valuable for students who are stuck, but at some point a fresh passage helps more than another repetition. The proposed rule above is "same passage on the first retry, new passage from the second retry on." Please confirm or adjust the cutover point.
"They can't read the passage. They can't access the passage that we gave them. So we want them to read that multiple times. So I think new words is important, but I think the same passage would be valuable for a lot of students that are stuck." Rebecca Patterson — 5/04 intro meeting, 1:03:00
"You could take the passage and have them read two sentences fluently, kind of dial it back. So you could use the existing passage to generate a smaller targeted practice out of it." Sarah Scott Frank — 5/04 intro meeting, 1:04:19
"Most of the time when the lessons get too difficult too quickly, especially in that CVC bucket, is because of the passage. Our passages start at lesson 12 and the kids that aren't ready to read a passage can't read a passage. So repeated exposure is definitely a good thing." Rebecca Patterson — 5/04 intro meeting, 1:04:38
The decision engine treats the group as the unit it reasons about — a group can be as small as one student. 1:1 tutoring is just a group of size 1, not a separate mode.
"We're building a tool that's for a group, a group that could be as small as one for one-to-one tutoring or, you know, up to four or five or six kids." Sarah Scott Frank — 5/04 intro meeting, 4:06
"This is a tool that's for groups, but your group can be as small as one. We specialize in one-to-one right now. What we do, we dabble in groups. And so it has to accommodate both." Sarah Scott Frank — 5/11 Sprint 1 Kickoff, 23:12
Each student has their own pathway underneath, but the group lesson stays at the slowest student's level. We're including this diagram so you can react to the actual behavior — is this the right model, or is there a better one your team has in mind?
"You really have to sort of go with what's best for the group, knowing there will be outliers. And the reality is that in lots of cases, the child can't move, like that's the third grade group… So you have to sort of teach to the mean and create a logic model that serves up something that's going to move a majority of kids." Sarah Scott Frank — 5/11 Sprint 1 Kickoff, 25:30
This is the pain we need to solve. We don't have a confident proposal yet — your team's input on the next iteration of group exit tickets would be valuable.
"I watched a lesson… there were seven kids in the group, maybe more. And every single, they sat there while every single student read aloud every single word on the slot for like nine minutes. It was a third of the lesson." Sarah Scott Frank — 5/04 intro meeting, 45:20
"If you have one, you can do X. If you have three, you can do X. If you have nine, you certainly cannot have nine children reading word lists of ten… So that becomes the whole lesson. That's dumb. So you can have a branching structure, or you can build a logic model that accounts for group size where something different happens for one child or even two children versus five." Sarah Scott Frank — 5/11 Sprint 1 Kickoff, 32:44
"Depending on the group size, there could be one at-bat per student, two at-bats per student, ideally three at-bats. Then they can at least take that and say, this group as a whole got 80% accuracy." Rebecca Patterson — 5/04 intro meeting, 34:44
We don't have firm thresholds for "most failed" yet — the right number depends on how you want the group to stay together vs. split.
Each section maps to a section of the question list we sent over:
When you're answering a question, look at the diagram to see exactly what part of the system your answer is shaping. The green "Where this came from" boxes under each diagram show the verbatim quote from your team that the proposal is based on — click the Fathom timestamp links to jump straight to that moment in the recording. Click any diagram to expand it to fill the screen for easier reading.
Click outside, press Esc, or use the × to close.