Show the full module text
Module 27 — Micro-Agreements
How an agreement is built in the session so that it survives an ordinary week at home.
This program has handed you a dozen things to take home, and all of them land in the same place: an ordinary week, in a house that was already full. This module is about how an agreement has to be built if it is going to survive that week. The way it is built is most of the work.
What this is about
Part Seven is where the program stops adding things and starts making them hold. Its three modules run in order. This one is how an agreement gets built. The next, Eight Dates, Neurodiverse, is a set of conversations. The last, Knowing It Is Working, is the one you open at about ninety days to see whether any of it worked.
Not at the door. An agreement said while somebody is putting a coat on is gone by Tuesday. It is built during the session, out loud, with both of you there, and written down where you will both see it.
When one does not happen, the sentence that arrives is we didn’t do it. Five different things hide inside that sentence. It was not remembered. It was remembered every day and would not start. Nobody knew what done meant. The week was a wall. Or one of you said yes because saying no is hard.
Four of those are design faults, and design faults are ours to fix.
Only the fifth is about whether either of you wanted to, and it is the rarest. So your therapist does not ask why didn’t you do it. Nobody can answer that accurately about themselves. The question is a location: where did it stop?
Once the five reasons are laid out as rows, most couples can point at theirs within a few seconds, and each row has its own fix.
What happens in the session
The agreement gets five protected minutes near the end of a session, not the last nine seconds. Your therapist builds one at a time, in front of both of you. It starts with a menu, not an instruction: three things the session could send you home with. Any one will do, and saying none of them fit is an ordinary answer.
Every agreement has the same six parts. One behavior. One person who starts it. One cue that already happens. One size, with a floor under it: the smallest version that still counts. One visible proof that it happened. And what happens if it does not. Your therapist fills each of them in out loud.
Filling all six takes about ninety seconds, and the one most often left empty is the last, because planning for a miss feels like inviting it.
The Agreement Session
1 What, exactly, and who starts it. A verb and an object: not connect more but ask her the three questions and write down what she says. Then one name. A task that belongs to both of you belongs to neither. If it is always the same partner starting things, that is the grievance, not the fix.
2 Attached to something that already happens. Not a time. An event: after the children are down, when the kettle goes on on Sunday, when the last laptop closes. A cue that lands in the middle of something absorbing is an interruption, and it will be resented and ignored.
3 How big, and what done looks like. Smaller than you think, with a floor underneath. And a finish line you can see: it is done when there are three things written on the card. Without that, a task gets abandoned or gets perfected.
4 What happens if it does not happen. Decided now, while nobody is ashamed. Here is the smaller version, and we would rather have that than nothing. If neither happens, come anyway and say so. That is information about the design, not a failed test.
Then one last question, which gives a polite no somewhere to go, followed by a silence that lasts longer than is comfortable.
On the cue. Not Tuesday. Tuesday needs somebody to notice it is Tuesday. What does it come after? After the kettle. After bedtime. After the laptop closes.
On the size. Halve it. Now halve it again. What is the version that still counts on the worst day of the week? That one is the agreement. The bigger one is there if you want it.
The real question. What would make you not do this? I am not testing you. I would rather find the problem now than in two weeks.
The answer to the last one, when it arrives, is usually something small and concrete that can be designed around in ten seconds.
Make it boring, then halve it
The size you first think of is sized for a couple whose week is elastic. So it gets cut to a quarter, and a floor goes under it.
The floor is the agreement, not a consolation prize.
A low bar that gets cleared leads a couple to ask for a higher one; a high bar that gets missed leads them to stop asking, and to stop saying so.
You can say no. Some people cannot do a thing when it is a requirement and can do the same thing when it is an option. So the agreement is offered as an option, and the offer is meant. Saying yes to get out of the room is not agreement.
The way this goes wrong. The agreement was that he texts at lunchtime on Wednesdays. He does it for two weeks and misses the third. Nothing is said. Four days later, in an argument about something else, it arrives: and you couldn’t even manage the text. Now the agreement is evidence, and he stops. The fix is a sentence said when the agreement is made, before anyone has anything to prove. This is not evidence. It does not get raised in an argument. If it did not happen, it comes back to the session and we fix the design.
Why we do it this way
Start with the reason that gets misread most: remembered every day, and it would not start. Autistic adults asked about this describe being unable to begin even simple things they fully intend to do, in a way that feels outside their control.1 What gets them moving is something from outside: a prompt, or a setting arranged so the action is already half made.
This is the state that gets filed as not caring, and the filing does more damage than the missed task.
It is also why the agreement is attached to an event, not a time.
Tuesday evening needs somebody to notice it is Tuesday evening. That is the exact capacity that is short.
After the kettle goes on needs nobody to notice anything.
Research on if-then plans, plans that say when this happens, I will do that, finds the same thing: the cue starts it, so the person does not have to.2
Why so small. New behaviors take far longer to become automatic than anyone expects. In one study of daily habits the range ran from a few weeks to most of a year, and missing a single day made no real difference.3 So a couple of weeks of feeling nothing is not evidence of anything. And a missed day is not a reset. That belief ends more agreements than any actual failure does.
Between-session work that gets done is linked to better results in therapy.4 It is easy to read that as pressing harder. It does not say that. It says that work a couple can complete does something, which makes the design of a completable agreement the intervention.
Where the rest comes from. The floor, the menu, the location question and the rule against keeping score do not come from a study. They come from what thousands of neurodiverse couples have told us. An agreement that becomes a scoreboard has stopped being an agreement.
After the session
Write the agreement where you will both see it without having to remember to look: the fridge, a reminder on the phone that already runs your week, the shared note that is already open. Our Externalized-Brain Setup worksheet is for deciding where, and is worth doing once for all of your agreements.
The check-in a week later. It does not open with did you do it, which puts one of you in front of a class. It opens with the design: I think I sized that badly. Then where did it stop. Then the redesign, handed to you: what would have worked?
An agreement the two of you rebuilt has a much better chance than one that was prescribed.
One honest limit. If nothing has happened for four weeks running and every session is spent redesigning, the design is not the problem. Something else is: a capacity crisis, a plan aimed at the wrong difficulty, or one partner who has quietly stepped out of the relationship and not said so. Your therapist will say that plainly rather than hand you a fifth version of the same sheet.
The workbook below is one agreement, built.
The one thing, if that is all you have. If this is too much this week, fill in two lines: who starts it, and what it comes after. Those two do more work than all the others together.
Your workbook
Your answers save to this device only - we cannot see a word of what you write. This one is a single agreement, built properly: the parts, the floor, where it lives, and the rule that keeps it from becoming a ledger.
The agreement
A verb and an object. If it contains an adjective about how you will be with each other, it is not there yet.
The agreement, in one line
Who starts it (a name)
It happens after ___ (something that already happens, not a time)
The floor: the smaller version that still counts on a bad week
How you will both know it happened
Where it is written down, where you will both see it
Before you agree
Answer these now, before there is anything to prove.
What would make you not do this?
We agree: this is not evidence and does not get raised in an argument. — Agreed, Agreed, and we have broken this rule before, Not yet - we should talk about that first
Where this comes from
Every source below was checked against the published record. They are grouped by the kind of evidence they are, so the numbers may not run straight down the page — a number points back to where the source is used in the lesson.
Research discussion
The tool in this module is a way of building an agreement in the session: one behavior, one name, one cue, one size with a floor, one visible proof, and a decision in advance about what happens if it does not happen. The five-way location question, the menu of three, the floor, the what-would-stop-you question and the rule against keeping score are practice moves, developed in use with neurodiverse couples and not tested as a package.
Four studies sit under it, and all of them are indirect. The inertia study is focus groups with thirty-two autistic adults: qualitative, self-selected, and describing experience rather than measuring it.1 It is the clearest account available of intention surviving intact while starting fails. The if-then planning meta-analysis pools ninety-four tests and finds a medium-to-large effect on goal attainment, but those tests are mostly individual goals in general populations, and none are two people agreeing something with each other.2 The lesson takes the mechanism and does not claim the number.
The habit study followed ninety-six volunteers doing a self-chosen behavior daily for twelve weeks.3 Its value is the spread rather than the median: a range so wide that no couple should measure themselves against a number. The finding that one missed opportunity did not materially affect the process is the one we quote most.
The homework meta-analysis pools seventeen studies of cognitive behavior therapy and is correlational: people who complete more between-session work do better, which is not the same as assigning more producing better outcomes.4 It is used here for one inference only, that completable work is worth designing for.
What no study shows is that building an agreement this way produces better outcomes for couples than describing one at the door. Nobody has run that trial. The claim the lesson makes is that the design of an agreement is the intervention, and that four of the five ways it fails are ours to fix rather than the couple’s.
Who this research was done with. The studies behind this module, none of them of neurodiverse couples, drew heavily on white, comparatively well-off, English-speaking participants. If your household carries pressures those samples did not — money, immigration, racism, disability, unsafe housing — the practice still applies, but the room you are practicing in is harder. That is the room, not you.
Peer-reviewed research
1. Buckle KL, Leadbitter K, Poliakoff E, Gowen E (2021) "No way out except from external intervention": First-hand accounts of autistic inertia. Frontiers in Psychology, 12, 631596. https://doi.org/10.3389/fpsyg.2021.631596 Six focus groups with 32 autistic adults aged 23 to 64 on difficulty starting, stopping and changing activities despite intention. Participants described being unable to initiate even simple wanted actions, in a way outside conscious control, with stress and mental health difficulties making it worse; external prompts and compatible environments restored action. Re-cited from Module 6A. Limitation: qualitative self-report from a self-selected group; the authors note some accounts suggest a movement-disorder component that was not tested.
2. Gollwitzer PM, Sheeran P (2006) Implementation intentions and goal achievement: A meta-analysis of effects and processes. Advances in Experimental Social Psychology, 38, 69-119. https://doi.org/10.1016/S0065-2601(06)38002-1 Meta-analysis of 94 independent tests of implementation intentions, if-then plans specifying when, where and how a goal will be pursued. They had a positive effect of medium-to-large magnitude (d = .65) on goal attainment, working by making the specified opportunity more accessible and automating the response to it. Re-cited from Module 6E. Limitation: largely individual goals in general-population samples; nothing about couples, households or neurodivergent participants, and the pooled estimate is from a 2006 synthesis rather than from recent better-controlled trials.
3. Lally P, van Jaarsveld CHM, Potts HWW, Wardle J (2010) How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998-1009. https://doi.org/10.1002/ejsp.674 96 volunteers chose an eating, drinking or activity behavior and performed it daily in a consistent context for 12 weeks, with 82 providing enough data for analysis. Nonlinear regressions fitted an asymptotic curve to each person's automaticity scores across 84 days; the time to reach 95 percent of that asymptote ranged from 18 to 254 days, and missing one opportunity to perform the behavior did not materially affect habit formation. Limitation: 96 volunteers performing simple individual behaviors; nothing about agreements between two people, and the widely quoted median is one point in a very wide range.
4. Kazantzis N, Whittington C, Zelencich L, Kyrios M, Norton PJ, Hofmann SG (2016) Quantity and quality of homework compliance: A meta-analysis of relations with outcome in cognitive behavior therapy. Behavior Therapy, 47(5), 755-772. https://doi.org/10.1016/j.beth.2016.05.002 Meta-analysis of 17 studies and 2,312 clients. Homework quantity related to post-treatment outcome with a pooled effect of g = 0.79 (15 studies, 1,537 clients) and quality with g = 0.78 (3 studies, 417 clients); associations persisted at follow-up. Limitation: correlational, so it cannot show that assigning more homework improves outcomes; the quality estimates rest on three studies with very wide confidence intervals, and none of the samples are couples in relationship therapy.
Further reading
• Neurodiverse Couples Counseling Center (2026) The Externalized-Brain Setup. Practice materials, Neurodiverse Couples Counseling Center. https://www.neurodiversecouplescounseling.com/worksheets/the-externalized-brain-setup The practice's worksheet for deciding where a thing lives: building capture systems so that working memory stops doing a job it cannot do, on the principle that out of sight really is out of mind. Used here for the cue and the place the agreement is written down. Limitation: a clinical tool developed in use rather than validated, and presented to couples on that basis.
The design of the agreement is the intervention
The Neurodiverse Couples Counseling Center works with couples where one or both partners are autistic, ADHD or AuDHD. We build the agreement in the session rather than describing it at the door: one behavior, one name, a cue that already exists, a floor for the worst week, and a rule that keeps it from turning into evidence. Therapy for clients in California, coaching worldwide, all by telehealth. A first conversation costs nothing.
Up next
Module 28 — Eight Dates, Neurodiverse