by Sprout’n
Harvest watches your stock against your real sales rate and your supplier's real lead time, then tells you when to order and whether the order leaves you short.
Your best product sells out on a Thursday.
The factory needs sixty days to make more and thirty to ship it. So the honest answer is that you lost the next ninety days of sales on that line, and you lost them two months ago, on the day you did not place an order you had no reason to think about.
The opposite failure is quieter and costs more. Forty thousand dollars sitting in a warehouse in a size that moves twice a month, bought because the minimum order quantity was round and the discount looked good.
Neither of those is a forecasting problem. They are both a timing problem, and timing is arithmetic once somebody is actually doing it.
Harvest reads sales from Shopify and from any channel you import, then works out how fast each product actually moves. The days a product was out of stock are removed from that calculation rather than counted as days nobody wanted it, because a naive average reads a stockout as collapsing demand and then tells you to order less.
Reorder by date equals stockout date minus lead time. Lead time comes from your own supply terms, split into production and freight because expediting shortens one of them and not the other. Where you have not entered terms, Harvest assumes ninety days and labels that row as an assumption rather than quietly using it.
Being late to reorder and being unable to pay for it are different problems, and the advice is opposite. Harvest keeps them separate. One says order now. Another says the deposit is more than you will have in the bank on the day it is due. A third says you can afford it but it takes you under the floor you set. A fourth says the money is already sitting in a warehouse.
Stock, supply and cash on a single axis. The gap in the top lane is the stockout. The bars below it are the production and freight that would have prevented it. The points along the bottom are when you pay for them. It answers when, which a cash chart cannot.
Worth saying plainly, because the limits are the reason to trust the numbers that are there.
Under four weeks of history, Harvest shows you what it has observed and no projection at all. Between four and twelve it gives a date and marks it provisional. When the uncertainty is wider than your lead time, it prints no date, because a window too wide to act on is worse than admitting the data will not support one.
A product that has never sold has no velocity. That is not a small number, it is no number, and the two must not look the same on a screen you are making a forty thousand dollar decision from.
Lead times, minimum orders and deposit terms come from you, because they come from a relationship rather than from an API. What Harvest does is make the consequences of them visible.
Harvest comes with everything
$25/month
for your first 6 months, then $50 a month
One subscription covers Roots, Harvest and Seed. Pay $300 for a full year inside the first 6 months and you keep the opening rate for all twelve.
Get startedHarvest is for the founder whose money is tied up in physical things, and who would like to know when it comes back.