Ask most store owners when they need to be ready for Black Friday and they will say Black Friday. Friday 27 November.
That is the wrong date, and planning against it is how stores end up making changes to a live theme while the traffic is arriving.
Any store run seriously stops touching its code about two weeks before the season. Not because Shopify says so, but because a theme edit shipped into peak traffic is the single most expensive mistake available to you: it breaks at the hour when every visitor costs money and nobody is at a desk. Two weeks is the buffer that lets a fault surface and get fixed while it is still cheap.
Which puts your real deadline at Friday 13 November.
From tomorrow, that is 55 Israeli working days. Fifteen of them sit inside the block between ראש השנה on Saturday 12 September and שמחת תורה on Saturday 3 October, and nobody who has run a business here plans those as full days. So call it 40 clean working days.
And if the plan is to start after the חגים, the number is 30.
Thirty working days for inventory, design, banners, campaigns, QA and any custom development. That is not a motivational statistic. It is what is left on the calendar.
The test: a date, or a plan?
Below are six questions. They are not a checklist of things to do, they are a readiness test, and each one has a pass condition that is deliberately strict.
You pass an item if you can answer it with a date and an artefact. A signed PO. An approved file. A booked scope. You fail it if the honest answer is an intention, however confident.
Most stores in late August will fail five of the six. That is normal and it is fixable. It is only a problem if you did not know.
1. Inventory: is the purchase order signed, or do you know what you would order?
Pass condition: the PO is signed and the supplier has confirmed the date.
Lead times of eight weeks and up are ordinary, and that is before a customs hold or a factory closure. This is the one item on the list that is already late in late August, and the only one where a late start cannot be rescued by working harder.
The version of this job most people get wrong: they open the sales-by-product report, sort by units sold, and reorder the top ten. That list is misleading, because the highest-volume product is very often a cheap entry item with almost no margin. Sort by net sales as well, and read the two lists together. The products high on both are where the money is, and those are the ones to over-order. High on units, low on net sales means a traffic driver: hold enough not to run out, no more. And anything that carried real revenue last November but sits outside your top twenty today deserves a second look, because a seasonal product looks weak in an August report by definition.
2. Custom development: is it scoped, or is it an idea?
Pass condition: the scope is written, the build is booked, and it lands before the freeze with a week of testing behind it.
This is the hardest deadline of the six and the one merchants misjudge most badly, because custom work is the item that feels most flexible and is in fact the least.
A bundle. Gift with purchase. Tiered discounts that unlock by cart value. A gift finder. A countdown that does not fight your theme. Every one of those is spec, then build, then test, and then it has to be inside the freeze. Working backwards from 13 November: specified by mid September, built through October, tested in the first week of November.
Which means the practical decision point for custom work is now, not October. If you commission bespoke development in November you are asking someone to write untested code into your highest-traffic week, and an agency worth hiring will tell you no. Being told no in November is much worse than deciding in September that you would rather not.
3. Design and banners: do the files exist, or does the brief exist?
Pass condition: every asset is designed, approved, and sized.
The failure here is almost never taste. It is volume, and the volume is consistently underestimated because nobody counts before briefing.
Count yours honestly. Homepage hero, desktop and mobile, which are different crops and not one file. Collection page banners for every collection in the sale. Product page badges. Email headers, one per send, and there is more than one send. Social in three sizes. Ad creative in several ratios. Then, if your store runs in Hebrew and English, double all of it, because a Hebrew banner is not an English banner with the text swapped: the composition flips.
That count is your schedule. Do it before you brief, because a designer given a list on 1 November and a list on 1 September is being asked two completely different questions.
4. Campaigns: are the audiences warm, or is the launch booked?
Pass condition: retargeting pools are filling now and the creative is approved before the freeze.
During BFCM week every advertiser in your category bids for the same attention in the same moment. What you pay that week is what it costs that week, and it is more than it costs today.
So the work that belongs in the next 40 days is the unglamorous half: filling retargeting audiences while attention is cheap, growing the email and SMS list, and getting creative approved early enough that approval is not sitting in the freeze week.
One thing worth changing while there is time. The default list-building offer is “10% off your first order”, and in the run-up to a discount season it works against you: you are training a list to wait for a discount three months before you plan to give them a much bigger one. Offer early access instead. Not a discount, a head start of a few hours on the sale. It costs nothing, it is genuinely valuable in November precisely because stock runs out, and it filters for the only trait that matters, which is someone who intends to buy from you rather than someone who wants a coupon.
5. QA: has someone actually bought something, or does it look fine?
Pass condition: a real person completed a real purchase with a real card on a real phone and received the confirmation email. On a staging copy, in the week before the freeze.
“It looks fine” is not QA. Every store carries a few quiet faults, and they are survivable at ordinary traffic and expensive at four times it, because the customers who hit them do not file a report. They leave.
Four checks, an afternoon:
Buy something yourself. Your own phone, cellular rather than wifi, all the way to a completed order and a confirmation email in the inbox. Not a test order from the admin. The real thing, refunded afterwards.
Request a shipping rate for every region you actually ship to, including the ones you sell to twice a year. A missing rate does not announce itself. It shows the customer nothing, and they conclude you do not deliver to them.
Run your best-selling product page, not your homepage, through a speed test with a slow connection simulated. The homepage is the page you optimised. The product page is where your ads land.
And open the checkout in a browser you never personally use. Most Israeli stores are built and tested in Chrome on a recent iPhone. A meaningful share of your buyers are on Safari, or on an Android two generations old, and that is where these faults live.
6. The freeze: do you have a date, or a general intention to be careful?
Pass condition: the date is set, everyone who can push a change knows it, and the exception rule is written down.
This is the item nobody has, and it is the one that makes the other five safe.
The rule that works is simple. After 13 November, prices, discounts, stock levels and campaign copy keep moving, because those are the levers you need during a sale. Code, theme, apps and integrations do not move, because those are the things that break stores.
Write the exception rule before you need it, not at 11pm on 24 November when someone wants to install one more app. In the moment it will always feel worth the risk. It is not.
What this does not mean
Three misreadings worth heading off.
It does not mean the חגים are wasted. Plenty gets done in that block. It means you should not schedule work inside it that needs four people and a decision.
It does not mean everything must be finished today. Only inventory and custom development are genuinely time-critical right now. The other four have real slack, which is exactly why counting is useful: it tells you where the slack is and where there is none.
And it does not mean that starting early guarantees a good season. There is a figure circulating this month that stores starting 90 days out earn two to three times the BFCM revenue of stores starting 30 days out. It appears only on the blogs of companies selling BFCM readiness services, with no study behind it, and we are not going to repeat it. Starting early does not make money. It removes the specific failures that lose it.
Six questions, forty days
Answer the six. Give yourself a pass only where there is a date and an artefact.
If you failed inventory or custom development, deal with those this week, because they are the two that a late start cannot rescue.
And put 13 November in the diary now, while it is still an ordinary date rather than an argument.
זהו להיום. נתראה בטיפ הבא של Shopping Dev Solutions.
