DT Cloud says cloud choice starts with critical services, what to protect first?

In his own words: “When choosing cloud infrastructure, we must first determine which services cannot stop.” So the question is not “which cloud” but “which service”. Change the order and the conversation changes too: a sales pitch turns into a needs meeting.

According to Aytulun, an organisation should know how long it can keep each service going if connectivity or management access is lost. The report does not suggest a time frame. You set it yourself: one for the till, one for orders, one for stock.

A critical service is a system whose failure stops your sales. In a store that means the POS and payments, in e-commerce the basket and checkout page, in the warehouse the order fulfilment screen. The report does not say which regulation defines “critical infrastructure”, so the term is used here in plain business language.

The moment you cannot take payment, the customer walks out and the basket is left hanging. Other systems can be run by hand for a while; payments usually cannot. So put a backup route for payments first, such as a second POS terminal or a second virtual POS provider for online sales.

When the order screen goes down, the sale is made but nothing ships. For the customer, that wait starts as patience and ends as a complaint. You need to know in advance who will work by hand on fulfilment, and from which list.

When the stock system stops, sales can go on, but you start risking selling products you no longer have. The longer the outage, the more errors build up SKU by SKU. If the latest stock data is regularly saved somewhere separate, recovery gets much easier.

If your stores are open, a short website outage does not end your sales. But if most of your revenue comes online, the website moves up to the same line as payments. The order is not fixed; it follows where your sales come from.

We suggest a simple tool: the outage card. Write one line for each critical service: the system’s name, what happens if it stops, how it runs by hand, who is responsible, who calls the provider. The card sits next to the till and on the manager’s phone. In a crisis nobody searches for documents; they open the card.

On store websites, the problem is often not the server but that nobody wrote down who does what. During an outage everyone looks at each other, and the work lands on whoever is nearest. A paper order form, a manual payment link, a confirmation call; these are ordinary tools, but they only work when they are written down.

Seeing that a backup was taken is reassuring, but restore tests often keep slipping down the calendar. A backup that has never been restored and tested is a hope. Servers do not go on holiday, but they love the week you do; run the test before that week.

E-commerce sites and store chains that write their critical services list in advance come out ahead. They know what they want when they talk to a provider and do not pay for capacity they do not need. In a separate statement dated 8 October 2026, Aytulun says today’s choice in cloud costs sets tomorrow’s bill.

Small online shops running their site and payments on a single cheap package with no backup will struggle. When an outage hits, they do not know who to call or what can run by hand. On your obligations for hosting payment and personal data, verify with the official source; this is not legal or financial advice.

Couriers, marketplace sellers and suppliers feel it indirectly. When a retailer’s order system stops, shipments stop too and the supplier’s delivery plan slips. When one link in the chain stops, the others wait.

An ad account does not always notice right away that your site is down, and the clicks keep coming. Add a line to your outage card: “who pauses the ads”. On a campaign day, that line matters more than your average basket.

Launches, sales or sponsorships that bring sudden attention push traffic well above normal. We look at how one brand draws attention through athletes in our piece on Bingo Jet Jel and athlete sponsorship. The infrastructure that has to meet that attention should be tested before the campaign.

As ad tests, customer lists and sales data grow, where they are stored becomes a decision of its own. We cover how a small brand sets up pre-launch ad testing in our article on Hayat Kimya’s testing lab. Test data deserves a line on the outage card too.

Yes, but it can start with a one-page plan. Writing down which system to protect first, what runs by hand during an outage and who is responsible is the first step for most small sites.

Look at your own critical services list before the price. According to DT Cloud CCO Aytulun, you should first decide which service cannot stop, then compare providers against that list.

Only by restoring it. Do not assume it works until you have restored it to a test environment and seen your orders and stock data open.