A&A INSIGHTS
When delivery fills the week: set a minimum contact volume
For a solo founder running delivery and sales: set a busy-week minimum contact volume, rebuilt by A&A from a YC AMA and Paul Graham.
日本語で読む
THE STARTING POINT
When sales stops in a week filled with delivery, the cause is not a lack of time but the absence of a rule for what may be stopped. Thinning everything pushes every activity below its information floor at once, so fully stop the work that would fall below the floor and keep only a single-channel minimum at full volume.
Founder sales stops for lack of a stopping rule, not for lack of time
A&A perspective
When sales stops in a week filled with delivery, it is not because there was no time but because you never decided what may be stopped. Without a rule, the decision itself costs time, so everything gets postponed together. Decide two things in advance: one minimum volume of contact you will execute even in a busy week, and a list of the sales work that may be stopped. Write both down on paper and the decision in a busy week takes ten seconds.
A&A perspective
The important part is that thinning everything is not the best move. Some sales work yields no readable result until it passes a certain volume. Five a week for twenty weeks and a hundred in one week are the same hundred but not the same information. In the first case, by the time responses arrive the premises have changed and you cannot decide what to repair. So what you do in a busy week is not reduce everything to a fifth; it is fully stop the work that would fall below its floor and keep one thing at full volume.
A&A perspective
This article reads the AMA with Gustaf Alströmer that Y Combinator published in 2018 alongside the essay Paul Graham wrote in 2013. From the first it takes a floor on contact volume and the identification of the buyer; from the second, the premise that the founder does the acquiring by hand, plus a caveat about porting this to a business that sells time. Source statements, A&A interpretation and hypothetical worked entries are kept distinct. Legal items to confirm when porting this to a Japanese service business are stated explicitly later.
The YC AMA: there is a floor below which contact volume yields no information
From the sources
In the AMA with Gustaf Alströmer that Y Combinator published on October 29, 2018, the way to think about sales is stated as: "you have to send at least 100 emails before you know if it works." The context is finding a scalable way to identify your audience and a scalable way to contact them, and the observation that email works for this at both small and large scale. It continues that you should just find 100 people you think are in your target audience and email them.
From the sources
The same AMA gives, as the procedure recommended to a B2B SaaS starting sales, first figuring out who the actual buyer is, the decision-maker who will pull out the credit card or sign the bill, and putting 100 matching LinkedIn profiles in a spreadsheet. The message is described as a plain-text email saying you are the founder and are reaching out to a small group of people you think would be a good fit for the product you are building, asking them to try it. As expectations it states that about 40% will open the email, 5-10% will click a link included, and less than 5% will try the product. Alströmer says the biggest mistake people make doing this is sending too few emails.
From the sources
The AMA also addresses who does this work. You should be your own first growth hire; growth in the early days is done by the founders, and because it is so core to the company it is not something you can outsource, similarly to how you cannot outsource product. On conferences, it states that conferences are not very scalable and that he has rarely seen it work. On the size of first customers, it says it is nearly always easier to start with smaller companies, ones that move fast and need your product but do not have a procurement team yet.
A&A perspective
What to take from this is the structure, not the figures. The count of 100 and the rates of 40% and 5% are a participant's view premised on US B2B SaaS in 2018 and are not a benchmark for Japanese service-business sales today. What can be taken is the property that contact volume has a floor below which results carry no information, and the consequence that an activity below its floor is evidence neither of failure nor of success. Thin everything in a busy week and every activity falls below its floor at once. The result is a week in which sales consumed time and taught you nothing.
| Sales work | Treatment in a busy week | Why it is judged that way |
|---|---|---|
| Status updates to existing clients and past enquirers | Do not stop (this is the minimum itself) | Response takes long, so the effect of stopping arrives as a trough two months later |
| The next move with a prospect who responded | Do not stop | Their deliberation is in motion, and a gap makes restarting expensive |
| Drafting a proposal with a deadline | Do not stop (prioritise if the deadline precedes delivery) | It is fixed to the other side's internal schedule and you cannot move it |
| Building and researching a new candidate list | May stop | The information does not degrade, and it can be done in a batch later |
| Writing new articles or posts | May stop (switch to reusing what is published) | Results take a long time, so one missing week has little effect |
| Sending an unsolicited proposal with no deadline | May stop | No expectation is in motion, and it competes directly with delivery quality |
| Attending meetups and trade shows | May stop | It consumes a long fixed block and the number of candidates gained cannot be predicted |
Graham's premise: acquisition is done by the founder, by hand
From the sources
In an essay dated July 2013, Paul Graham writes that "the most common unscalable thing founders have to do at the start" is to recruit users manually. Nearly all startups have to; you cannot wait for users to come to you, and you have to go out and get them. He also writes that for a startup to succeed, at least one founder, usually the CEO, will have to spend a lot of time on sales and marketing.
From the sources
On measuring progress, the essay recommends that every startup measure their progress by weekly growth rate, with the example that if you have 100 users you need to get 10 more next week to grow 10% a week. It also says that Airbnb was so fragile early on that about 30 days of going out and engaging in person with users made the difference between success and failure. Among early tactics that usually do not work it names the Big Launch and partnerships with large companies.
From the sources
The essay also carries a caveat that bears directly on service businesses. Consulting is the canonical example of work that does not scale, but it is safe to do so long as you are not being paid to, and that is where companies cross the line. So long as you are a product company merely being extra attentive to a customer, they are very grateful even if you do not solve all their problems. But "when they start paying you specifically for that attentiveness" — when they start paying you by the hour — they expect you to do everything.
A&A perspective
This caveat is the branch point when a service business borrows overseas advice. The part about going out to acquire by hand does port: the founder owns sales, contacts a small number of people individually, and learns by meeting them directly. That holds regardless of the shape of the business. The part about using attentiveness as a free learning mechanism does not carry the same meaning in a business that sells time. Attentive treatment of a customer mid-delivery is already the object of payment, so it does not become a structure that yields learning about new customers. In other words, in project work, generous delivery cannot substitute for sales. This is A&A's reading.
Setting the minimum, and the list of work that may be stopped
A&A perspective
Write the minimum as three things: one channel, a count per week, and a cut-off time within that week. The point is not to have more than one channel. Running two or more at full volume in a busy week is impossible, and the result is that both fall below their floors. Choose the count from your recent record on whichever channel gets you the most response, taking a floor you can comfortably fill in one week. Fix the cut-off time so that the order in which delivery overruns eat the sales slot is decided in advance.
A&A perspective
Two tests decide whether something may be stopped. First, does the information degrade if it is interrupted? Researching a candidate list or writing a new article will read the same next week. Second, is there an expectation already in motion on the other side? For a prospect who responded, or a proposal request with a deadline, a gap stalls their deliberation and restarting is expensive. Split on these two and you find that more work may be stopped than you expected, and less work may not.
A&A perspective
The table below is A&A's tabulation of that test. What the sources establish is the view that contact volume has a floor, the procedure of identifying the buyer and building a list of 100, the view that conferences are not very scalable, the claim that founders own growth, and the premise that the founder acquires by hand. The may-stop and may-not-stop assignments on each row are A&A's design, to be tested in the reader's own business.
A hypothetical example: allocating a week with two deliveries colliding
Hypothetical example
The following is a hypothetical example for explanation only; it is neither a real customer nor A&A results. Suppose a founder taking on AI workflow design work alone has two deliveries landing in the same week, with four hours available for sales. Previously those four hours were split one hour each across candidate research, writing an article, contacting existing customers, and drafting a proposal. All four ended half-done, and nothing was learned from that week's sales.
Hypothetical example
Fixing the minimum first changes the allocation. Narrow the channel to one, a status update to people with whom there has been a prior engagement or enquiry, set the count at eight a week, and the cut-off at 18:00 Thursday. Two of the four hours go there; the remaining two go to replying to the two prospects who responded and confirming their next meeting dates. Candidate research, writing a new article, and any proposal without a deadline stop entirely that week. If delivery overruns into Friday, the eight by 18:00 Thursday are already done.
A&A perspective
The point of this example is that total time did not fall. It stays at four hours; only the allocation changed. What changed is whether a readable result remains at the end of the week. If one of eight replies, next week you change the wording or the target. Even zero is information, because it is a zero that met a floor of eight. One hour each across four activities cannot support the same judgment. The counts and hours here are illustrative placeholders, to be reset from your own record.
Items to confirm when porting this to a Japanese service business
A&A perspective
This article does not recommend sending unsolicited sales email to an undefined audience. The list of 100 described in the source AMA is a method premised on US B2B SaaS in 2018. Taking the same shape in Japan brings in the requirements for advertising email under the Act on Regulation of Transmission of Specified Electronic Mail and comparable rules, the route by which personal data was obtained and its stated purpose of use, and the handling of contact details originating from business cards or enquiries. These are items for the reader to confirm for their own situation; this article is not legal advice. Until that confirmation is done, restrict the scope to existing clients, people who have enquired before, and people who have responded.
A&A perspective
That constraint actually fits the minimum-volume design well. Restricting scope to existing relationships keeps the total audience in the tens or low hundreds. The count that meets a floor therefore becomes naturally small, landing on realistic numbers like eight or twelve a week. There is no need to adopt the source's 100 as a target. What is needed is to touch, in one week, a share of your own audience large enough for the result to be readable. This is A&A's interpretation.
A&A perspective
The other constraint is the time-selling property seen in the previous section. In project work, attentive treatment of a customer mid-delivery is the performance of the contract itself, so increasing it does not produce learning about new customers. If anything, the added attentiveness tends to become a precedent you cannot charge for in the next estimate. The sales minimum has to be secured outside delivery; it cannot be made up for with delivery quality. That is what Graham's caveat means read from the service-business side.
Where this design does not fit, and what is not being claimed
A&A perspective
At a stage where referrals alone fill the pipeline, this design is unnecessary. If the referrers are a fixed handful and the relationships are stable, the object of the minimum becomes maintaining those relationships; rewrite it as a frequency rather than a count. Likewise, where the backlog runs six months out and negotiating higher rates matters more, a design for increasing contact volume can wait. What this article addresses is the state of repeatedly losing inbound interest once delivery finishes.
A&A perspective
This article does not claim that fixing a minimum contact volume wins engagements. Both sources are views grounded in the participants' experience, not measured success factors. The design of a weekly minimum is itself an A&A hypothesis, and no A&A close rate or engagement record is presented. Nor does it recommend using the source's open or trial rates as your own forecast. Measure again on your own channel, against your own audience.
A&A perspective
Finally, a note on the side effect of stopping what may be stopped. Pause writing new articles for a few weeks and search-driven traffic falls, but late: not in the week you stopped, but two or three months later. So stopping is fine, but record the weeks you stopped. When you later look at a change in traffic, you can attribute the cause correctly to the pause in publishing rather than to the sales channel. Graham's recommendation to measure progress weekly is useful for tracking exactly this kind of delayed causation.
The next step: two lines to write this week
A&A perspective
Two lines this week. Line one, the minimum: "channel is X, N per week, cut-off is day D at time T." Line two, at least three pieces of work that may be stopped, named. Do not write "sales activity" in the abstract; write specific names like "candidate list research," "writing new articles," "proposals without a deadline." Named, they can be stopped in a busy week without guilt. The following week, add the count executed and the count that responded to the same sheet.
A&A perspective
If you want the whole picture from acquisition through retention first, see A&A's guide "AI-native GTM: a practical guide for solo founders and small teams." How to build the mechanism that keeps buyer-specific proposals alive during delivery is covered in "Turn solo-founder sales into a weekly system: keep buyer-specific reasoning alive while shipping," and the minimum this article sets is meant to sit on top of that mechanism. If the judgment is split on what the minimum should be, we can also set it together in a consultation, bringing your recent response record.
The busy-week decision shrinks to ten seconds once the minimum and the stoppable list are written down in advance. Write the minimum as one channel, a count per week, and a cut-off time, and name at least three pieces of work that may be stopped. The source's 100 emails and open rates are views premised on US B2B SaaS; in Japan, restrict the scope to existing relationships and people who have responded, and reset the count after confirming the legal requirements for advertising email yourself.
Sources & editorial note
Primary pages read for this article. Publication dates below belong to the sources; access dates record our research.
- Growth AMA with YC Partner Gustaf Alstromer
Y Combinator · 2018-10-29
Accessed 2026-10-02 - Do Things that Don't Scale
Paul Graham · 2013-07
Accessed 2026-10-02
AI-assisted editorial production
A&A uses AI for research, writing, translation and editorial checks. Source facts, our analysis and hypothetical examples are labeled separately.
Editorial check: 2026-10-02