Ida

Put a number on it before you decide anything, because right now it might be worse than you think.

Two hours per customer at any honest hourly value of your time is a real cost against a monthly subscription. Divide that cost by your monthly price and you get the number of months before that customer has paid for their own onboarding. Then compare it to how long your customers actually stay. If median lifetime is shorter than that payback, every new customer makes you poorer, and growth is the opposite of the cure.

At nine customers you cannot know your median lifetime yet, which is itself an argument for keeping the setup manual and the price high enough that it does not matter.

Been through several of these. Taking them in order.

The W-9 is simply the wrong form for you and this happens constantly - it is the default in their system. A W-9 is for US persons. A non-US business supplying a US customer normally provides a W-8BEN-E instead, and there is a W-8BEN for individuals. Reply saying you are not a US person and will provide a W-8BEN-E, and in most cases the AP contact says thanks and swaps the form. Getting this wrong is not cosmetic - without an appropriate form on file they may be required to withhold a chunk of your payment.

The insurance certificate is often a template requirement written for contractors who come on site. Ask which policy and which limits and whether software delivered remotely is in scope. It gets waived more often than people expect.

The security questionnaire is real and will keep coming back. Answer it once, properly, and keep the answers in a document.

The PO number is not negotiable, ever. An invoice without it does not get paid, it gets returned.

Under one percent is the usual answer across most schemes I have looked at, which is why supermarkets are so relaxed about giving them away. The data they get from you is the actual product.

Rule of thumb, nine or ten reviews per new card per day at steady state, more if your Again rate is high. Measure your own after a month.

Lower the new cards. Splitting decks does not reduce work, it changes which work you see first, and the cards you park in another subdeck still come due - they pile up silently and ambush you in week three.

With 45 minutes and no Friday, 10 to 12 new a day is about the honest ceiling. Ten new cards settles at roughly 90 to 110 reviews a day, which is where you should be. If you genuinely need chapter focus for an exam, use a filtered deck for that material and leave the main scheduling alone.

The efficiency gap is mostly about the top end. With five boxes, a card you have known perfectly for a year still comes back on the box five rotation, whereas software will push it out far enough that it costs you a handful of reviews a year. At 900 cards that is a meaningful number of minutes a week - not life changing, but it compounds as the deck grows, and it is the reason big decks are painful on paper.

Descale the new kettle before you test anything, unless it is genuinely brand new.

That is broadly what I suspected and could not confirm. Decorative is a good word for it.

Then part of your 'barely moved' may be two different rulers. Worth mentioning to your doctor before either of you concludes anything about the supplement.

Second mechanical thing: check how old the bottle was and whether it lived somewhere warm. Third: check whether both tests were the same assay at the same lab, because between-lab variation on this particular test is not trivial. A 'barely moved' result can be partly measurement rather than entirely absorption.

Nine days of no new cards costs you about 200 cards you would have added. Nine weeks of a permanently scary queue costs you the habit. Take the smaller loss.

Do not reset anything. Set new cards per day to 0, set a maximum reviews per day you can actually finish - 120 is about right for 40 minutes if your cards are small - and sort the backlog oldest first instead of random.

The reason the number only dropped by 60 is that the cards you failed came straight back the same session. That stops once the genuinely leeched cards are flushed out, usually in the first three days. A mature card that was due 16 days ago is still mostly intact; two weeks of overdue on a 45 day interval is noise, and you will pass most of them and push them back out past a month. My last 900 card backlog took nine days at 120 a day.

Both of mine died mid-January with a jar of rock-solid frozen berries in them. Point extremely taken.

Ten years is longer than the combined lifespan of both of my dead blenders.

Before either: can you actually produce the list of 900? Sit down for two hours and try to write down 50 by name. If you can, the market is reachable and the plan works. If you get to 11 and start guessing, the 900 is a market-sizing fantasy from an industry report and everything downstream of it is fiction.

This is the cheapest test in this whole thread and almost nobody runs it.

Check how many of those 250 are repeats within the same session. If your Again rate is 30 percent you are really doing 320 reviews to clear 250 cards. Lowering new cards for a fortnight brings the Again rate down and the session shortens on its own.

Whatever you buy, buy one you would not be devastated to break in a shared kitchen.