---
title: "Digital transformation in fashion: what research shows"
description: "What two open-access reviews of fashion's digital shift count and miss, plus the EU ecodesign duties facing small brands and makers. As of 3 Nov 2026."
url: "https://staging.khoi.co.uk/en/blog/so-hoa-nganh-thoi-trang-ban-do-nghien-cuu"
lang: en-GB
author: "Lê Quốc Khôi"
date: 2026-11-03
updated: 2026-11-03
topics: ["Supply chain & operations"]
tags: ["Digital transformation", "Fashion industry", "Literature review", "Digital product passport", "EU ESPR"]
---

# Digital transformation in fashion: what two research reviews count, and what they miss

## Key takeaways

- Two open-access reviews count topics and methods; they do not measure effects. Sarker and Bartok (2024) keep 46 papers from 518 records (Sarker and Bartok, 2024, pp. 1, 6); Nam and Kim (2025) analyse 249 articles from Korean journals (Nam and Kim, 2025, Methods). Sarker and Bartok themselves list the effectiveness of individual technologies as a gap (Sarker and Bartok, 2024, p. 17).
- Three reviews, three numbers: 46, 249 and 224 papers, the third from the abstract of Liu and Liu (2025) (Liu and Liu, 2025). The first two used their own search terms and scope, and I do not know the third's query, so the numbers cannot be read as the growth of one field.
- In the Korean sample, the keyword AI selected 23 of 316 articles, against 79 for e-commerce, 52 for platforms and 46 for the metaverse (Nam and Kim, 2025, Methods).
- EU law is more specific than the research: the Ecodesign for Sustainable Products Regulation (ESPR), Regulation (EU) 2024/1781, requires a digital product passport for product groups covered by a delegated act (European Parliament and Council of the European Union, 2024, Art. 9(1)), and the Commission's plan gives 2027 as the indicative year for textiles (European Commission, 2025).
- Since 19 July 2026, Article 25 bans destroying unsold consumer products listed in Annex VII, except for micro and small enterprises; medium-sized firms follow from 19 July 2030 (European Parliament and Council of the European Union, 2024, Art. 25(1)). Annex VII lists apparel, clothing accessories and footwear (European Parliament and Council of the European Union, 2024, Annex VII); Recital 57 gives the reason (European Parliament and Council of the European Union, 2024, Recital 57).

One review counts 224 journal articles on digital transformation in fashion between 2011 and 2024 (Liu and Liu, 2025). Another keeps 46, and a third 249. All three describe the same industry, so which number says how far the research has got?

This post is for the owner of a small fashion brand or garment workshop being pitched 3D, RFID (radio-frequency identification), AI or the metaverse. I read the running text of two open-access reviews and set it beside three official EU documents; Liu and Liu I could read only as an abstract. Everything is as of 3 November 2026.

## What do the two open reviews count?

Topics, methods and papers per year; not the effect of any technology. The table sets the two side by side, with Liu and Liu at abstract level.

> **Figure:** (compare) Table compiled by me from the texts of the three papers; the "Cannot say" row is my judgement after reading.

```json
{
  "type": "compare",
  "columns": ["Sarker and Bartok (2024)", "Nam and Kim (2025)", "Liu and Liu (2025), abstract only"],
  "rows": [
    { "label": "Question", "cells": ["Which green and digital transition factors appear in research on the fashion industry?", "How has Korean research on fashion's digital shift changed over three periods, across five factors: business model, process, service, product, customer?", "Update the research picture and propose future directions"] },
    { "label": "Search", "cells": ["Scopus, Web of Science, ScienceDirect, Google Scholar; \"green transition\" or \"digital transition\" with \"fashion industry\" or \"textile industry\"; last searched 18 March 2023", "Korea Citation Index journals only; ten keywords chosen after scanning Naver news from November 2022 to November 2023", "Unknown beyond the abstract"] },
    { "label": "Sample", "cells": ["518 records, 46 papers kept (2012–2022)", "316 articles by keyword; 249 once sorted into five factors (2014–2023)", "224 articles (2011–2024)"] },
    { "label": "Reported result", "cells": ["Eight digital factors: IoT (Internet of Things), AI, robotics, 3D printing, blockchain, AR (augmented reality), social media, online shopping", "Customer 132 articles, service 58, product 29, business model 15, process 15; peak in 2022", "Three intellectual clusters, six thematic keyword clusters; the abstract names focal areas such as sustainability, AI-enhanced luxury experiences and consumer engagement in digital retail"] },
    { "label": "Cannot say", "cells": ["The text I read gives no measured effect of any technology", "Uses titles, abstracts and keywords only; Korean journals only; I did not open the 14 tables", "I read nothing beyond the abstract"] }
  ]
}
```

Sarker and Bartok screened the titles and abstracts of 473 records, excluded 391, took 82 forward, and kept 46 papers (Sarker and Bartok, 2024, pp. 1, 6). The figures 2.17%, 6.52%, 8.69%, 10.86%, 28.26% and 43.47%, which the paper calls annual growth rates for 2017 to 2022 (Sarker and Bartok, 2024, p. 8), add up to almost 100%, so they are shares of the 46 papers. In counts that is 1, 3, 4, 5, 13 and 20 (my arithmetic), which matches the paper's own statement that 2022 had 20. In this sample the count rises, from a very small base.

Nam and Kim selected articles with ten keywords: e-commerce 79, platform 52, metaverse 46, game 39, virtual reality 28, big data 25, AI 23, mobile app 14, chatbot 8, cloud 2, 316 in all; after two fashion-marketing experts classified them, 249 remained (Nam and Kim, 2025, Methods). Output peaked in 2022 with 74 articles, with 59 in 2023 (Nam and Kim, 2025, Results). They report 159 qualitative against 121 quantitative studies (Nam and Kim, 2025, Results).

## Why do three reviews give three numbers?

Because each asks and searches differently, the numbers cannot be added or compared. Sarker and Bartok required "transition" together with "fashion" or "textile", and admit the search terms could be broader (Sarker and Bartok, 2024, p. 16). Nam and Kim took only Korean journals, and their ten sampling keywords came from one year of news coverage (Nam and Kim, 2025, Methods). Liu and Liu count 224 articles, but I do not know their query (Liu and Liu, 2025).

The consequence is that a topic's rank in a review reflects the review's own keyword list. In the Korean sample, "metaverse" and "chatbot" are in because Korean news mentioned them often; "RFID", "3D" and "blockchain" are not among the ten sampling keywords, though RFID and 3D still appear among the process-factor keywords (Nam and Kim, 2025, Results). I infer that a paper about RFID alone may have missed the sample. That is my inference, unchecked.

## What do the reviews say about how well each technology works?

In the parts I read, almost nothing. Sarker and Bartok's conclusion says digital technologies are used "to optimise production efficiency, decrease costs, and raise product quality" (Sarker and Bartok, 2024, p. 16), but that summarises the papers they cite and comes with no measurement of their own. The claim that 3D printing made physical sampling "obsolete" likewise rests on other papers' citations (Sarker and Bartok, 2024, p. 15).

They list eight research gaps themselves, three of which bear directly on an owner's question: environmental benefits need quantifying (Sarker and Bartok, 2024, p. 16); enterprise-level empirical work combining expert opinion and surveys is needed (Sarker and Bartok, 2024, p. 16); and studies of the effectiveness of specific technologies, such as blockchain for supply chains, 3D printing for production and virtual reality for retail, are still missing (Sarker and Bartok, 2024, p. 17).

Nam and Kim measure something else: which keywords sit next to which in titles, abstracts and keywords, and they read a keyword's "centrality" as the influence of the topic (Nam and Kim, 2025, Results). That measures attention, not profit or cost. To see what an effect figure looks like without a baseline, read [Burberry's +24%](https://staging.khoi.co.uk/en/blog/burberry-cong-nghe-xoay-chuyen); for a company assessing its own RFID, read [Inditex's unified stock](https://staging.khoi.co.uk/en/blog/omnichannel-ton-kho-thong-nhat).

## Where do small firms and Vietnam sit on this map?

Almost off it. Among the 23 countries of authors' origin across the 46 papers, China has 7, Italy 6, the UK 4 and India 4, and Vietnam is not named (Sarker and Bartok, 2024, p. 9). The paper mentions Vietnam once, in a sentence about adoption trends in India, Bangladesh, Vietnam and Ethiopia that cites two other sources (Sarker and Bartok, 2024, p. 4). Nam and Kim cover Korea only and say so (Nam and Kim, 2025, Limitations). No paper I read measures small firms, and I have not searched Vietnamese-language literature.

## What does EU law require that the research does not?

A deadline and a kind of data to record, independent of any evidence on effectiveness. Regulation (EU) 2024/1781, the Ecodesign for Sustainable Products Regulation (ESPR), says a product covered by a delegated act may be placed on the market only if a digital product passport is available under that act (European Parliament and Council of the European Union, 2024, Art. 9(1)); the Commission may exempt a product group from this requirement (European Parliament and Council of the European Union, 2024, Art. 9(4)). The Commission's 2025–2030 working plan ranks textiles, apparel in particular, first among four final products in the ranking by the Commission's Joint Research Centre (JRC), with 2027 as the indicative adoption year; footwear is kept apart, awaiting a study due by the end of 2027 (European Commission, 2025, sections 2.2.1, 2.2.3).

According to the text, the passport is a record linked through a data carrier to a persistent unique product identifier (European Parliament and Council of the European Union, 2024, Art. 10(1)). Each delegated act will set the data, the carrier, and whether the passport sits at model, batch or item level (European Parliament and Council of the European Union, 2024, Art. 9(2)). Manufacturers must ensure it is available (European Parliament and Council of the European Union, 2024, Art. 27(1)), and so must importers (European Parliament and Council of the European Union, 2024, Art. 29(2)). The working plan says imports onto the EU market will have to meet the new standards too (European Commission, 2025, section 3.1) and that the passport will include material composition and substances of concern (European Commission, 2025, section 3.2). Recital 33 says that where the passport relies on standards that are not free of charge, the impact assessment should consider whether such reliance is appropriate and how disproportionate costs for micro, small and medium-sized enterprises can be avoided (European Parliament and Council of the European Union, 2024, Recital 33).

The rule on unsold goods already applies. Since 19 July 2026, Article 25(1) prohibits destroying the unsold consumer products listed in Annex VII, except for micro and small enterprises; medium-sized enterprises follow from 19 July 2030 (European Parliament and Council of the European Union, 2024, Art. 25(1)). Annex VII lists apparel, clothing accessories and footwear by commodity code (European Parliament and Council of the European Union, 2024, Annex VII); Recital 57 gives the reason (European Parliament and Council of the European Union, 2024, Recital 57). "Destruction" includes recycling, other recovery and disposal; preparing for reuse does not count (European Parliament and Council of the European Union, 2024, Recital 55). Article 24 requires an annual public disclosure of the number and weight of unsold products discarded, the reasons and where they went, again except for micro and small enterprises, with medium-sized enterprises following from 19 July 2030 (European Parliament and Council of the European Union, 2024, Art. 24(1)).

Delegated Regulation (EU) 2026/296, applying from 19 July 2026, sets ten derogations, such as dangerous goods, goods infringing intellectual property rights and damaged goods that cannot be repaired; firms must keep evidence for five years (European Commission, 2026, Arts. 2–3). Where there is enough evidence that micro and small enterprises could be used to get round the rules, the Commission may extend the ban and the disclosure duty to them (European Parliament and Council of the European Union, 2024, Art. 25(5)).

What I take from this: the passport and the unsold-stock disclosure both ask for data (codes, batches, composition, stock left over and where it went), and neither requires a specific tagging technology, at least in the text I read. That is my inference.

## How should a small brand use these sources?

To choose questions, not technologies. Suppose (a hypothetical of mine) you run a 30-person workshop in Binh Duong, selling T-shirts through an importer in Germany. The last column is my reasoning, not a recommendation from the authors or the Commission.

| Question | Anchor | What I would do |
|:--|:--|:--|
| Does this technology have before-and-after figures for a firm my size? | Effectiveness of each technology is a stated gap (Sarker and Bartok, 2024, p. 17) | Ask the seller for figures and a baseline; if none, run a small trial with a comparison group |
| What data will my importer need, and when? | Manufacturers and importers both must ensure the passport exists (European Parliament and Council of the European Union, 2024, Art. 27(1); European Parliament and Council of the European Union, 2024, Art. 29(2)) | Ask them directly; the data list for clothing does not exist yet |
| Does each style and batch have one lasting code, and where does unsold stock go? | The passport is tied to a unique identifier; Article 24 asks for numbers, weight and destination (European Parliament and Council of the European Union, 2024, Art. 10(1); European Parliament and Council of the European Union, 2024, Art. 24(1)) | Fix style and batch codes, keep material composition by batch in one place, and log leftover stock even if you may be exempt, before choosing any tagging technology |
| Should I buy AI, 3D or metaverse tools first? | The keyword AI selected 23 of 316 articles in the Korean sample; topic frequency is not effectiveness (Nam and Kim, 2025, Methods) | Pick by one measurable operating question, not by how hot the keyword is |

This suits you if you sell or are about to sell into the EU, or need a reason to decline a pitch. It does not suit you if you sell only at home with no export plan: EU law does not yet touch you, and two reviews are not enough to choose technology.

## Going deeper: where are these reviews weak?

Sarker and Bartok start in 2012 because "the first digital technologies were introduced to the industry in 2011", citing two other papers (Sarker and Bartok, 2024, p. 5). They own their limits: only articles and reviews, no books or conference papers, and the search could add "apparel industry" or "transformation" (Sarker and Bartok, 2024, p. 16). I see two slips: the eight green factors in Table 3 do not quite match the paragraph just after it (the paragraph has "green building", the table does not) (Sarker and Bartok, 2024, pp. 10–11), and the text does not explain the gap between 518 and 473 records; Figure 3 is an image I could not read from the text layer.

Nam and Kim say the study offers "universal insights" into how digital transformation works in fashion, usable globally (Nam and Kim, 2025, Conclusions), then concede that the data come only from Korean journals and need comparison with international data (Nam and Kim, 2025, Limitations). The counts are also loose: 316 articles in the final analysis, then 249 by factor, and the text does not say where the other 67 went; nor does it say whether 159 qualitative and 121 quantitative studies are counted out of 316 or 249 (Nam and Kim, 2025, Methods; Nam and Kim, 2025, Results).

Both look only at published texts, so unpublished work and firms' internal data fall outside the sample (my inference about a limit common to this type of review). I use no industry reports such as McKinsey's or Business of Fashion's.

## What did I read?

Sarker and Bartok (2024): the journal's full PDF (21 pages, open access), including the methods, Tables 1–4 and the list of gaps; I did not open the papers they cite. Page numbers are the PDF's printed ones.

Nam and Kim (2025): the full open-access HTML, from introduction to limitations. I did not open the 14 tables (only "full size table" links), so the figures I give come from the running text. The HTML has no page numbers, so I give section names.

Liu and Liu (2025): abstract only, via Crossref; the article is closed and there is no open copy.

Regulation (EU) 2024/1781: recitals (1)–(38), (49) and (55)–(57), Articles 9–10, 23–26, 27, 29 and 79–80, and Annexes III and VII; I did not read all 2,300-odd lines of the body, but I searched the whole text, annexes included, for "RFID" and "blockchain". The working plan COM(2025) 187 and Delegated Regulation (EU) 2026/296: in full. General limits: this post says nothing about the business results of any technology, about Vietnamese firms or Vietnamese law, and I have not checked the Official Journal after 3 November 2026.

## Frequently asked questions

### How does a systematic review differ from a bibliometric analysis?

Sarker and Bartok (2024) run a systematic review under PRISMA 2020, the standard reporting guideline for such reviews, then a descriptive and content analysis (Sarker and Bartok, 2024, p. 1). Nam and Kim (2025) use bibliometric analysis, which quantitatively examines the literature to identify trends, plus keyword frequency and network analysis (Nam and Kim, 2025, Introduction). Liu and Liu (2025) combine the two: their abstract describes a systematic review using bibliometric and thematic analyses (Liu and Liu, 2025).

### Does the digital product passport require RFID or blockchain?

Neither word appears anywhere in the Regulation, annexes included. Article 10 requires the passport to be linked through a data carrier to a persistent unique product identifier, with data based on open standards and, as appropriate, machine-readable (European Parliament and Council of the European Union, 2024, Art. 10(1)). Recital 37 gives a watermark or QR code as examples of a data carrier (European Parliament and Council of the European Union, 2024, Recital 37). Annex III points to the ISO/IEC 15459 family of standards for identifiers and data carriers (European Parliament and Council of the European Union, 2024, Annex III).

### Is there a free copy of Liu and Liu (2025)?

I checked Unpaywall on 3 November 2026: the article is listed as closed access with no repository copy. The publisher's page blocks automated access, so I read only the abstract, through Crossref (Liu and Liu, 2025). This post therefore uses none of Liu and Liu's detailed results.

### When does the digital product passport start to apply to clothing?

I have not found a date. The Commission's April 2025 working plan gives only an indicative 2027 for adopting the textiles delegated act (European Commission, 2025). Article 9(4) also lets the Commission exempt a product group from the passport requirement where technical specifications are not yet available (European Parliament and Council of the European Union, 2024, Art. 9(4)). I have not checked the Official Journal after 3 November 2026.

## References

- European Commission (2025) *Ecodesign for sustainable products and energy labelling working plan 2025-2030 (COM(2025) 187 final)*. (Communication from the Commission, Brussels, 16 April 2025). Available at: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52025DC0187 (Accessed: 2026-11-03).
- European Commission (2026) *Commission Delegated Regulation (EU) 2026/296 of 9 February 2026 supplementing Regulation (EU) 2024/1781 by setting out derogations from the prohibition of destruction of unsold consumer products*. (Official Journal of the European Union, L, 22 April 2026). Available at: https://eur-lex.europa.eu/eli/reg_del/2026/296/oj/eng (Accessed: 2026-11-03).
- European Parliament and Council of the European Union (2024) *Regulation (EU) 2024/1781 of the European Parliament and of the Council of 13 June 2024 establishing a framework for the setting of ecodesign requirements for sustainable products, amending Directive (EU) 2020/1828 and Regulation (EU) 2023/1542 and repealing Directive 2009/125/EC*. (Official Journal of the European Union, L, 28 June 2024). Available at: https://eur-lex.europa.eu/eli/reg/2024/1781/oj (Accessed: 2026-11-03).
- Liu, S. and Liu, C. (2025) ‘Mapping the digital transformation in the fashion industry: the past, present and future’, *Journal of Fashion Marketing and Management*, 29(7), pp. 1292–1319. https://doi.org/10.1108/JFMM-09-2024-0380
- Nam, S. and Kim, S. (2025) ‘Deciphering the digital evolution: strategic insights into the Korean fashion industry through advanced analytical perspectives’, *Fashion and Textiles*, 12(1), article 6. https://doi.org/10.1186/s40691-025-00415-0
- Sarker, S.I. and Bartok, I. (2024) ‘A systematic review of green and digital transitional factors in the fashion industry’, *Business Systems Research*, 15(1), pp. 1–21. https://doi.org/10.2478/bsrj-2024-0001

---

Canonical HTML version: https://staging.khoi.co.uk/en/blog/so-hoa-nganh-thoi-trang-ban-do-nghien-cuu