This material is for owners of online stores and companies with a large catalog who are launching a new site or moving the range to another one. If there are about ten products, they can be entered by hand without rushing. If there are hundreds of them, the way the data is prepared starts to affect the deadlines and the estimate more than the design of the card.
I will go through what information is needed for each product, how to arrange the table, how to match photographs to the cards, and why tidy data makes it possible to load the catalog automatically.
What data is needed for each product
The set depends on the range, but the basis repeats in almost any store.
- Name.The one the buyer will see, not an internal warehouse designation.
- Article number.A unique code by which the product can be found without ambiguity. It links the table row, the photo and the card on the site.
- Category.The catalog section the product will sit in, and if there are several sections — all of them.
- Price.And, if there is one, the old price for a discount.
- Specifications.Size, material, color, weight, volume — everything the buyer chooses and compares by.
- Variants.If one product is sold in several sizes or colors, these are variants of one product, not different products.
- Description.Short or detailed, but about the specific product.
- Photographs.One main one and, if needed, additional ones.
Stock and availability are worth thinking about separately. If they change often, it is important to decide at once where the site will take them from and who will update them.
How to arrange the product table
The main rule: one row — one product or one product variant, one column — one property. As soon as one cell holds the size, the color and a comment at once, automatic loading stops working, and such cells have to be taken apart by hand.
Inside a column, the values must be written the same way. “Cotton”, “cotton 100 %” and “cot.” are the same thing to a person, and for a filter on the site they are three different values. The same with units of measurement: if one row has “50 cm” and in another “0.5 m”, comparing the products becomes difficult.
Category names are also worth writing the same way in all rows. A typo in a section name turns into an extra catalog section.
It is convenient to keep the table in Google Sheets or Excel. What matters is that there is one current version, not several copies with different employees.
An example table for importing products
Below is a notional example. This is a scheme, not the format of a specific platform: the exact column names for import depend on the system, and before loading I check them against its help.
| Article number | Name | Category | Price | Material | Color | Size | Photo |
|---|---|---|---|---|---|---|---|
| TS-001 | Basic T-shirt | T-shirts | 1 200 | Cotton | White | M | TS-001-1.jpg |
| TS-001 | Basic T-shirt | T-shirts | 1 200 | Cotton | White | L | TS-001-1.jpg |
| TS-002 | T-shirt with a pocket | T-shirts | 1 500 | Cotton | Gray | M | TS-002-1.jpg; TS-002-2.jpg |
In the example the first T-shirt has two sizes, and these are two rows with one article number: that way the system will understand that these are variants of one product. The prices are notional. If there are no variants, one row is enough for each product.
How to match photographs to the cards
The most reliable way — to name the files by the article number. The main photo — “TS-001-1.jpg”, the additional ones — “TS-001-2.jpg”, “TS-001-3.jpg”. Then the file name shows which product it belongs to, and in the table it is enough to list the names.
If photos of different colors differ, the color or the variant should be added to the name as well: “TS-001-white-1.jpg”. Otherwise, when choosing a color, the buyer will see the wrong shot.
What works worst is when the photos lie apart from the data and without labels. Then each card has to be assembled by looking at the shot and guessing which article number it is. On a large catalog this is long manual work, and it is easy to make a mistake in it.
The shots for the cards themselves are worth thinking about even before the shoot: the same background, angle and proportions. Bringing hundreds of dissimilar photographs to one look takes a long time.
Why structured data speeds up loading
When the data lies in a table by the rules above, the catalog can be loaded by import: the system reads the file and creates the cards itself. For an online store on Tilda these are CSV and YML files. Manual work remains for checking and fixing individual errors.
If the data is scattered, import does not help: first it has to be collected, taken apart and brought into a table, and it is this part that takes the main time. A substantial part of the effort in a store is often tied exactly to product content: collecting the data, searching for the source files, selecting the shots, bringing the photographs to one look and bulk loading.
The structure pays off after launch as well. If the range is updated, a new table is loaded the same way, and not entered again one item at a time. Changing prices for a group of products is also simpler in the table than in each card separately.
How to check the table before loading
Before handing the table over, it is worth looking through it quickly. Most errors are visible without special tools if the columns are sorted one by one.
- Whether different products have the same article numbers.
- Whether prices and categories are filled in on all rows.
- Whether the names of categories, colors, materials and units of measurement are written the same way.
- Whether all the files listed in the photo column are in the folder, and whether the names match down to the character.
- Whether there are extra spaces, line breaks and comments such as “check with the manager” in the cells.
The last point seems minor, but such a comment easily ends up on the site together with the data. Notes are better moved to a separate column that will not be loaded.
If the data is still in different places
Products are often described in several sources: a price list with the manager, specifications in the supplier’s catalog, photos in folders by the date of the shoot, some of the descriptions on the old site. This is a normal starting point.
Everything can be gathered into one table on your side if the company has a person who knows the range. If there is no time for that, I can take this work on: at the start I can fill the catalog myself from your materials, and what I cannot determine I will check with you. The volume of such preparation is included in the estimate in advance.
You can also start with part of the range. Sometimes it is reasonable to launch the store with the main categories and add the rest as the data is prepared.
Questions and answers
Do all the specifications need to be filled in right away?
No, but the ones by which the buyer chooses and filters products are better filled in from the very start. If a filter by size appears later, a size will have to be added to all the products after the fact. Secondary specifications can be added after launch.
Who will add the products after launch?
Whichever is more convenient. Products can be added in the site’s control panel on your own, or this work can be given to me. If the updates are regular, it is convenient to include them in site support: catalog filling and price updates are part of it.
What should be done with descriptions if there are none?
For the start, short descriptions with the main specifications are enough, and the detailed texts can be written gradually, beginning with the best-selling products. Copying descriptions from supplier or competitor sites is better avoided: such texts do not set your store apart from others.
If you want to estimate how much work your catalog in particular will take, tell me about it in the quiz on the site. I will look at the form the data is in now and suggest where to start.