Online shop on WooCommerce

It's the theme and thirty plugins that make it slow, not the name of the engine. We measured two of our own WordPress sites on July 31, 2026: one had a first server response of 285.9ms with 119.5KB of HTML, the other had a first response time of 1,520.2ms with 224.6KB. The difference is 5.3 times on the same core and the same constructor. Therefore, at the start, we put one cache instead of three, keep the list of plugins short and fix the target response time in the TOR, not after complaints. Fair margin: speed removes technical loss, not creates demand - if the price is higher than the market, a fast site will simply show it to the buyer faster.

A few hundred to a few thousand items work smoothly with normal hosting and cache. The technical reason for the limit is this: the product is stored as a record in wp_posts, and the price and attributes are stored in the wp_postmeta meta table, so the filter by characteristic becomes a join on that table. Woo removes part of it with lookup tables, and it is enough for several thousand records. For twenty thousand SKUs with variations, separate engineering begins - a separate search index, a batch import, a revision of the database indexes - and then it is fairer to count either another engine or a large catalog as a separate task.

Let's say it like it is. The stores we measured on July 31, 2026 run on OpenCart; WordPress in the measured portfolio is two service sites and a native application. That's why WordPress numbers are above, not "Woo store cases", which we didn't measure. Separately, about why we measure at all: in the card of one of our projects, the stack was signed as WordPress/WooCommerce, and the external measurement showed OpenCart 3 - we corrected the case description, instead of leaving a more convenient version. If you need a contractor with a dozen stores listed on Woo, that's a normal requirement, and we won't pretend that it's closed to us.

Yes, this is a typical task. We transfer the content and structure manually or with a script, we fill in the goods from export or from the supplier's price, we close the old addresses with 301 redirects according to the correspondence map. Rescheduling adds 5-10 business days to the deadline. Honest about expectations: redirects save addresses and transfer weight, but do not save items - after the move, the delivery is rebuilt, and for the first weeks it is visible in Search Console.

Yes, through the official API: the buyer chooses the city and branch or post office from the list that is pulled up in real time, and the TTN is formed from the same data simply from the admin. The benefit should be counted as savings on order processing, and not as an increase in traffic: the manager stops specifying the address in the correspondence and typing it into the carrier's office by hand. And a fair limit — this saving turns into money only if the freed time is really redistributed; integration does not reduce the cost of the delivery itself and does not speed up the carrier.

The WordPress core itself is updated, but plugins are usually broken - that's why we give the list of installed versions and the rule: first staging, then updates, then running the checkout with a test order. We close one thing even before launch: by default, the open REST API returns /wp-json/wp/v2/users with a list of logins, that is, a ready list of targets for sorting passwords. Continuous support is not included in the development and is taken separately; during the warranty period, we correct errors from our side, and not the consequences of independent changes in the code by third parties.

Because the amount is based on a figure that is not visible from the outside: how many entries your catalog will give with variations. Three hundred products of eight sizes is 2,400 records, and in terms of import time and filter behavior it is a completely different project than three hundred products without variations. Therefore, first, a free audit in 2–4 working days, and according to its figures, the amount and term by stages. The document with the numbers remains yours regardless of who will do the shop.

No one guarantees a position, and we will not. We close the technical part: CNC, canonical on the filter combination, site map announced in robots, Product and Offer markup. There are two limits that should be known in advance. The first is that the site map does not increase the position and does not guarantee indexation, this is the wording of Google itself. The second is that the markup does not raise the position either, it only makes the page suitable for the extended view, and the non-existent markup is downright dangerous: in our own application, we found JSON-LD with a rating of 4.8 and 50 reviews with zero reviews on the page, and this is a reason for manual sanctions, not an advantage.