In a fashion shop, the size selector frequently decides whether a parcel stays with the customer or comes back. Anyone who sells trousers, jackets or shoes online knows the pattern: the same size label fits differently depending on brand and cut, customers order two sizes to be on the safe side, and part of the goods is returned. This article shows how to build size advice without promotional promises: with size charts by brand and cut, a fit scale derived from recorded return reasons and clean size attributes in the product feed. The figures come from returns research, from surveys and from the rules that matter for anyone planning online shop development in fashion. They do not promise that any single measure has an effect, and neither does this article.
Why size decides the way back
How large the share of fashion in the volume of returns is can be seen in the analysis by the Research Group on Returns Management at the University of Bamberg for the data year 2021. Almost one in four parcels in German online retail went back to the retailer, a total of 24.2 percent of parcels (European Return-o-Meter, University of Bamberg). An estimated almost 530 million return shipments were transported (European Return-o-Meter). Fashion accounts for the largest part of this: 83 percent of returned shipments and 91 percent of returned items were fashion (European Return-o-Meter). For 2025, the research group expected approx. 550 million return parcels, based on a return rate assumed to remain unchanged at 24.2 percent (Research Group on Returns Management). That is a projection and not a count. For fashion shops, returns are therefore not a side issue.
Why goods come back was the subject of a Bitkom survey in autumn 2023. Two thirds (67 percent) of respondents cite a size that did not fit as a reason for returns (Bitkom). The question referred to products in general, such as clothing or furniture, with multiple answers allowed; the value is therefore neither a share of all returns nor a figure for fashion only. In the same survey, 29 percent of respondents state that they have deliberately ordered more than they needed at least once, for example clothing in different sizes (Bitkom). Overall, online shoppers in Germany send back 11 percent of their online purchases by their own account, across all product types (Bitkom). Size is thus one of the points a shop can address itself: it cannot influence how a fabric drapes, but it can decide what information customers receive before they click on the basket.
None of the sources cited measures how much a size chart reduces wrong sizes or returns, and none puts a figure on the effect of a single measure. The survey values describe reasons and habits of the respondents, the return figures describe the market as a whole. What size advice achieves in your own shop only becomes visible in your own data, and even there only if return reasons are recorded cleanly.
One estimate of the savings potential for the overall market comes from an older publication. In December 2019, the Research Group on Returns Management estimated that functioning online size advice and binding, standardized size information from manufacturers could together save up to 25 percent of returns in the overall market, around 120 million items (University of Bamberg). This figure applies to both measures together, not to a size chart alone, and it is an estimate, not a measured effect. Because the research group has changed its evaluation method, it cannot be compared directly with the figures for the data year 2021. Anyone quoting it should keep the "up to" and should not derive a calculation for their own shop from it.
Body measurement or garment measurement: what the standards describe
A frequent reason for contradictory size information is that two kinds of measurement are confused. The European standard DIN EN 13402-3:2017-12 describes the size designation of clothes on the basis of body measurements and intervals (DIN Media). It expressly does not specify measurements for garments (DIN Media). How the body measurements themselves are taken, that is where the tape measure sits at the chest, waist or hips, is defined in DIN EN ISO 8559-1, German edition 2021-04, which is based on ISO 8559-1:2017 (DIN Media). In practice this means: a size designation says which body a garment is intended for. How wide the garment itself turns out depends on the cut, the material and the ease that a brand builds in.
| Feature | Body measurement (according to DIN EN 13402-3) | Garment measurement (size chart for the item) |
|---|---|---|
| What is measured | The body the size designation is intended for | The garment laid flat, for example the chest width from seam to seam |
| Basis | Body measurements and intervals | Pattern, material and the ease built in by the brand |
| Information on the garment | No measurements for garments | Exactly these measurements, per size and cut |
| Method of measuring | Defined in DIN EN ISO 8559-1 | To be defined by the shop and described in measuring instructions |
| Use in the shop | Indicates which body a size is intended for | Comparison with a well-fitting garment from your own wardrobe |
For the shop, this results in a clear division of labor. The body measurement chart answers the question of which size fits one's own body, the garment measurement chart the question of how the specific item turns out. Many customers prefer to measure a favourite garment that fits well rather than measuring themselves with a tape measure. Size advice should therefore offer both routes. How this work fits into an overarching concept for returns is described in our article on returns management in e-commerce.
The standards mentioned describe designations and methods of measurement; this article does not derive any obligation for shops from them. They are useful as a common language: if you align your measuring instructions with DIN EN ISO 8559-1, you describe body measurements the way manufacturers and standardization bodies understand them, and you make information from different sources easier to compare.
One data model per brand and cut
A size chart that applies to the entire range is easy to maintain, but it answers the wrong question. A slim-fit pair of jeans and a relaxed-fit pair of jeans from the same brand carry the same size label and still have different waist and thigh widths. Between two brands, the deviations can be even larger. A sound model therefore assigns measurements neither to the individual product nor to the category, but to the combination of brand, cut and size. A product refers to its cut, the cut carries the size chart, and each variant refers to one row of this chart.
- Brand: manufacturer or own label, because every brand uses its own basic patterns and intervals.
- Cut: for example slim, regular or relaxed, each with its own size chart instead of one chart per category.
- Size designation: the value customers see in the size selector, together with the size system in which it applies.
- Garment measurements: the values measured on the garment laid flat, per size, for example chest width, shoulder width, length, waist width or inside leg length.
- Body measurements: the range of body measurements a size is intended for, if the brand supplies them.
- Measuring instructions: a reference to the description of how the shop measures, so that values from different deliveries remain comparable.
# Size chart for one cut, maintained per brand
brand: "Own label"
cut: "regular"
category: "shirt"
size_system: "EU"
measured_on: "garment" # garment measurement, not body measurement
unit: "cm"
guide: "measuring-guide-tops"
sizes:
- label: "S"
chest_width: null # chest width, flat from seam to seam
shoulder_width: null # shoulder width
length: null # length from the highest shoulder point
- label: "M"
chest_width: null
shoulder_width: null
length: null
body_range: # only fill in if the brand supplies body measurements
source: "manufacturer"
values: nullWhere this data lives depends on the system landscape. In many fashion shops, measurements come from supplier spreadsheets, from the ERP system or from a product database. A PIM system connected to the shop is an obvious place, because brand, cut and variant attributes are maintained there anyway. It is important that sizes are set up as genuine variant attributes and not as free text in the product title, because only then can they be used equally in the size selector, the size chart, the returns analysis and the feed. How variants can be structured cleanly without search engines seeing duplicate content is shown in our article on product variants and duplicate content.
Size charts are only as comparable as the measurement behind them. Define whether measurements are taken flat or on the body, where the tape measure starts, whether seams are included and how values are rounded. Photograph the measuring points once per product type and use the images internally and in the shop.
Keep size systems and size types clearly apart
A size is not unambiguous without its system. The same number can denote different sizes in different countries, and this is particularly visible with shoes. In the data model, the size system therefore belongs to every size designation as a separate field. The Google Merchant Center Help describes what happens otherwise: if the size system attribute [size_system] is missing, Google applies the standard system of the target country, and the size information can be distorted as a result (Google Merchant Center Help). The situation is similar with size types such as plus size, petite or maternity. For items with two size types, for example big and tall, two values can be submitted in the size type attribute [size_type] (Google Merchant Center Help).
A fit scale built from return reasons
The second pillar of size advice is knowledge of how a cut is actually worn. This information is contained in the return reasons, provided they are recorded in a structured way. In practice, this is frequently where the gap lies: only 29.5 percent of the online retailers surveyed record return reasons in a fully automated way, and for 34.4 percent manual recording is still the basis of data collection (EHI Retail Institute). If reasons are collected as free text on a slip of paper or in a comment field, they can hardly be analyzed by brand, cut and size. The first step is therefore not a tool but a fixed vocabulary of reasons that is used in the same way in the returns form, in the warehouse and in customer service.
- Define reasons: a short, fixed vocabulary such as "too small", "too large", "too tight in one place", "too long", "too short" and "fit different than expected", supplemented by reasons unrelated to size.
- Record reasons where they arise: in the returns portal as a mandatory selection with optional free text, and in the warehouse as a check of whether the stated reason matches the goods.
- Link every return to the brand, cut and size of the variant, not just to the order number.
- Count size-related reasons per cut and size and set a minimum number below which no statement is shown.
- Derive a tendency from the ratio of "too small" to "too large" and show it in a few levels, for example "runs slightly small", "true to size" and "runs slightly large".
- Recalculate the tendency at fixed intervals and record changes to the cut or the size chart with a date, so that old and new data are not mixed.
The fit scale does not replace the size chart; it complements it from the customer's point of view. Together, they answer the question that is actually asked at the size selector: do I take my usual size or one size up? It is important that the scale only evaluates returns with a size-related reason and leaves all others aside. A jacket that comes back because of its color says nothing about its fit. The following query shows the principle using a table of recorded return items.
-- Fit tendency per brand, cut and size from recorded return reasons
SELECT
r.brand,
r.cut,
r.size_label,
COUNT(*) AS size_returns,
SUM(CASE WHEN r.reason = 'too_small' THEN 1 ELSE 0 END) AS too_small,
SUM(CASE WHEN r.reason = 'too_large' THEN 1 ELSE 0 END) AS too_large,
1.0 * (SUM(CASE WHEN r.reason = 'too_large' THEN 1 ELSE 0 END)
- SUM(CASE WHEN r.reason = 'too_small' THEN 1 ELSE 0 END))
/ NULLIF(COUNT(*), 0) AS fit_score -- below zero: runs slightly small
FROM return_items r
WHERE r.reason_group = 'size'
AND r.returned_at >= :since
GROUP BY r.brand, r.cut, r.size_label
HAVING COUNT(*) >= :min_returns;The scale shows a direction that emerges from what customers report. It depends on who orders, how reasons are recorded and how many returns are available. So do not show decimal figures in the shop, but a few understandable levels, and hide the scale as long as the minimum number has not been reached.
Not every return of a second size is an advice problem. According to a Bitkom survey from autumn 2023, 29 percent of respondents have deliberately ordered more than they needed at least once, for example clothing in different sizes (Bitkom). Such selection orders can be identified in the data when the same order contains several sizes of the same item. They belong in a separate analysis so that they do not distort the fit scale.
Where the size chart and the scale sit on the product page
A good size chart is of little use if it is hidden behind a small text link at the bottom of the page. Size advice belongs where the decision is made: right next to the size selector. There, the entry to the size chart and the note on fit should be visible without scrolling, on a smartphone as well. Which other elements belong in the first visible area of a product page is described in our article on product page optimization above the fold. For size advice, four principles have proven themselves that apply regardless of the shop system.
At the size selector
The entry to the size chart sits directly at the size selector and not in a tab further down the page.
One chart per cut
The chart shown is the one for the cut the product belongs to, instead of a general chart for the whole category.
Fit in words
The tendency appears as a short sentence, for example "runs slightly small", and only once a defined minimum number is reached.
Units and systems
Centimeters, inches where they suit the target group, and the size system are visible in every chart.
On a smartphone, a wide size chart quickly becomes unreadable. What usually reads well is a genuine HTML table with a fixed first column for the sizes, horizontal scrolling for the measurements and a highlight for the size currently selected. Images of tables cannot be read by screen readers; an HTML table with column headers and a caption can, and it can be generated from the same data as the feed. The overlay or expandable panel should open and close with the keyboard and return focus to the size selector after closing.
- Size chart per brand and cut instead of one chart per category
- Label showing whether body or garment measurements are displayed
- Measuring instructions with an image for every product type
- Size system and unit visible in every chart
- Fit note only once a defined minimum number of returns is reached
- Chart as HTML with column headers, not as an image
- Entry to size advice next to the size selector without scrolling
- Size attributes in the feed from the same data as the display in the shop
Customers' body data: sparing and voluntary
Some size advisers ask for height, weight or body type and store the answers in the customer account. That can be convenient, but it involves personal data that many people regard as sensitive. Anyone collecting such information should keep it voluntary, state the purpose clearly, store only what is necessary and make deletion easy. The advice should also work without this information, for example by comparison with a garment of one's own. How we handle data on our own website is set out in the XICTRON privacy policy; clarify the requirements for your shop with your legal adviser or your data protection officer.
Size attributes in Google Merchant Center
What is modeled cleanly in the shop should also arrive in the product feed. In Google Merchant Center, the size attribute [size] is required for all products in the categories clothing (ID 1604) and shoes (ID 187) (Google Merchant Center Help). Since May 2026 there has additionally been the variant option attribute [variant_option]; Google recommends submitting size as a variant-defining property there and additionally in the [size] attribute (Google Merchant Center Help). For a fashion shop with many variants this means: size must be available per variant as a separate field and must not be extracted from titles or free text only at export. The following overview maps the four attributes to the fields of the data model.
| Attribute | Purpose (according to Google Merchant Center Help) | Source in the data model |
|---|---|---|
| [size] | Required for clothing (ID 1604) and shoes (ID 187) | Size designation of the variant |
| [size_system] | If missing, the standard system of the target country applies and the size can be distorted | Size system per size chart |
| [size_type] | For two size types, for example big and tall, two values are possible | Size type per cut |
| [variant_option] | Since May 2026; Google recommends size there and additionally in [size] | Variant-defining attributes of the product |
This mapping is manual work if sizes are maintained as free text in the shop, and largely automatic if they come from the data model. The handover is taken care of by a Google Merchant Center integration that reads size, size system and size type per variant from the same fields that feed the size selector in the shop. Shop and ad then show the same size, and discrepancies surface during the feed check instead of only at the customer.
What a return costs and where it goes
For the costs of a return, the European Return-o-Meter gives mid values: per returned item, transport and processing costs were 2.85 euros, per return shipment 6.95 euros (European Return-o-Meter). Which kind of mid value is meant is not stated in the public summary. As a mid value, 93.2 percent of returned items could be resold as good as new (European Return-o-Meter). Even so, every return causes shipping, inspection, refurbishment and restocking. With seasonal fashion, an item that is only available again after the end of the season may only be sold at a discount.
The return conditions give an indication of the general setting. In the German subsample of the European Return-o-Meter, 88.7 percent of the retailers surveyed offered their customers free returns, in the rest-of-EU subsample the figure was 52.4 percent (European Return-o-Meter, data year 2021). The size of the subsamples is not published, so the comparison only shows a direction. For size advice this means: where returns cost customers nothing, the threshold for ordering several sizes is low, and information that creates clarity before the order becomes all the more important.
How varied the situation in fashion retail is can be seen in a survey by the EHI Retail Institute. In the textile trade, roughly one in eight retailers surveyed (12.2 percent) gets more than half of their orders back; the remaining 87.7 percent report return rates of up to 50 percent (EHI Retail Institute). These are shares of the retailers surveyed, not shares of returns. The majority of respondents do not expect a quick improvement: a good 80 percent of the online retailers surveyed do not expect the volume of returns to improve over the next three years (EHI Retail Institute). Size advice is therefore more of an ongoing task of range maintenance than a one-off project.
The environmental side, too, has only partly been measured. Fewer than 5 percent of the retailer representatives surveyed for the European Return-o-Meter stated that their company measures the carbon footprint of returns (European Return-o-Meter). Among these retailers, the mean value was approx. 1,500 g CO₂e per returned shipment (European Return-o-Meter). Taking this value as a basis, returns in Germany caused an estimated 795,000 tonnes of CO₂e in 2021 (European Return-o-Meter). This is a scenario calculation by the research group and not a measured value for the market. It is based on the small group of companies that measure at all.
Ecodesign Regulation: destruction ban for clothing and footwear
Since 19 July 2026, the destruction of unsold consumer products listed in Annex VII of the Ecodesign Regulation has been prohibited; Annex VII lists apparel and clothing accessories as well as footwear (Regulation (EU) 2024/1781). The ban initially applies to large enterprises. It applies to medium-sized enterprises from 19 July 2030, and micro and small enterprises are exempt (Regulation (EU) 2024/1781, Art. 25(1)). One definition matters for fashion shops: under Art. 2(37), products returned by consumers under their right of withdrawal also count as unsold consumer products (Regulation (EU) 2024/1781). For clothing and footwear from Annex VII, the destruction ban therefore in principle also covers returns.
Exceptions are laid down in Delegated Regulation (EU) 2026/296. Under it, apparel, clothing accessories and footwear may exceptionally be destroyed if, owing to damage, deterioration or contamination including hygiene defects, they can reasonably no longer be used by consumers, even where the consumer caused the damage, and if repair and refurbishment are not technically feasible or not cost-efficient; the documents under Art. 3 must be available on request (Delegated Regulation (EU) 2026/296). For context: the share of returns that online retailers disposed of themselves was 1.3 percent in 2021 (European Return-o-Meter). The regulation concerns destruction, not the return itself. A link to size advice is only indirect: a wrong size that is not ordered in the first place does not have to be inspected, refurbished or disposed of. That is a conclusion of this article, not a statement of the regulation.
This section summarizes the legal acts according to their wording and does not replace legal advice. Whether and from when the ban affects your company depends on its size, and which exception applies in an individual case is something to clarify with your legal adviser. For documentation, it helps if return reasons, condition of the goods and their whereabouts are recorded in a structured way per item, which is the same data the fit scale needs.
Who buys fashion online
The target group for size advice is large. 48 percent of people aged 16 to 74 in Germany ordered clothing, shoes or accessories online in the three months before the survey in 2024 (Destatis). 53 percent of women and 44 percent of men in this age group bought clothing online in the same period (Destatis). There are also differences in sending goods back: by their own account, women return 14 percent of their online purchases and men 9 percent (Bitkom). These values apply to all product types and are self-reported.
Another figure comes from a different survey: 68 percent of online shoppers bought fashion and accessories online in the past twelve months; according to Bitkom, this is the product category bought online most frequently (Bitkom). This figure refers to online shoppers and a period of twelve months, whereas the Destatis figure refers to all people aged 16 to 74 and three months. The two values can therefore neither be offset against each other nor compared.
Implementation with XICTRON
Size advice is not a single plugin but an interplay of data model, returns process, product page and feed. We build these parts into existing shops or take them into account from the outset as part of online shop development. The starting point is usually an inventory: where are measurements kept today, how are return reasons recorded, and which size attributes arrive in the feed? A first overview of load time, search engine signals and accessibility of your shop is provided by the free XICTRON shop check; we then discuss size advice specifically on the basis of your range.
- Data model: brand, cut, size and measurements as structured fields, connected to the ERP system or PIM.
- Size charts: output per cut as an accessible HTML table with measuring instructions and a choice of unit.
- Return reasons: fixed vocabulary in the returns portal and in the warehouse, linked to the respective variant.
- Fit scale: analysis per brand, cut and size with a minimum number and understandable levels.
- Feed: size, size system, size type and variant option from the same data as in the shop.
- Maintenance: processes for new cuts, supplier data and corrections so that the charts stay up to date.
Size advice touches on several topics we have covered this week and in earlier articles. If you want measurements and fit notes to be answered in a chat as well, the article on shop chatbots with maintained knowledge explains the prerequisites for keeping answers up to date. If you are combining changes to the product page and data model with a version change, you will find pointers in the article on preparing for the upgrade to Shopware 6.8. And how returns can be predicted from order data is described in the article on AI return prediction.
Would you like to know how far your product data already goes for size advice? Send us a short description of your range, your shop system and the place where measurements are maintained today. You can reach us directly via the contact form for your size advice enquiry; we will get back to you with an assessment of which steps make sense in your case and in which order they can be implemented.
This article draws on data and texts from the following sources: European Return-o-Meter of the Research Group on Returns Management at the Otto-Friedrich-Universität Bamberg (data year 2021) and its projection for 2025, a press release of the University of Bamberg on measures against returns, Bitkom (survey on online shopping and study report "Digitaler Handel in Deutschland"), the Federal Statistical Office (Destatis), EHI Retail Institute, DIN Media (catalog entries for DIN EN 13402-3 and DIN EN ISO 8559-1), Google Merchant Center Help as well as Regulation (EU) 2024/1781 and Delegated Regulation (EU) 2026/296. Figures may vary depending on the time of data collection.
A body measurement chart assigns to a size the body measurements it is intended for, such as chest, waist and hip girth. A garment measurement chart describes the item itself, for example the chest width from seam to seam or the length. DIN EN 13402-3 bases size designation on body measurements and intervals and expressly does not specify measurements for garments (DIN Media). For a fashion shop, both charts are useful, provided it is clearly labeled which one is being shown.
As a rule, no. Cuts from the same brand turn out differently, and between brands the deviations can be even larger. A sound approach is one chart per combination of brand and cut, to which each product refers. New products of a known cut take over the chart without being measured again, and corrections apply to all products of that cut.
First define a short, fixed vocabulary of reasons and introduce it in the returns portal and in the warehouse. Older free-text entries can often be roughly assigned to the new reasons, but they are usually too imprecise as a basis for a fit scale. A fresh analysis from the day of the changeover makes sense. In parallel, it is worth looking at the technical condition of the shop; the free shop check for load time, SEO and accessibility provides a first overview.
There is no reliable figure for an individual shop. In December 2019, the Research Group on Returns Management estimated that functioning online size advice and binding, standardized size information from manufacturers could together save up to 25 percent of returns in the overall market (University of Bamberg). That is an estimate for both measures together and not a measured effect of a size chart. What is possible in your own shop only becomes apparent in your own, cleanly recorded return reasons.
For clothing and shoes, the size attribute [size] is required (Google Merchant Center Help). The size system [size_system] should be supplied as well, because otherwise Google applies the standard system of the target country and the size information can be distorted as a result (Google Merchant Center Help). Depending on the range, size type [size_type] and variant option [variant_option] are added. It makes sense to generate all values from the same fields the shop uses.
For large enterprises, the destruction of unsold apparel, clothing accessories and footwear has been prohibited since 19 July 2026; for medium-sized enterprises the ban applies from 19 July 2030, and micro and small enterprises are exempt (Regulation (EU) 2024/1781). Under Art. 2(37), products returned under the right of withdrawal also count as unsold consumer products (Regulation (EU) 2024/1781). Exceptions, for example for damaged goods, are laid down in Delegated Regulation (EU) 2026/296. Checking the individual case is a matter for legal advice.