The Translation and Localization Industry in India: Survey Results Released

印度的翻译和本地化行业: 调查结果发布

2020-10-05 12:39 slator

本文共4903个字,阅读需50分钟

阅读模式 切换至中文

The Slator 2020 Indian Language Service Provider (LSP) survey was conducted by Slator in collaboration with India’s language industry association, CITLoB (Confederation of Interpreting, Translation and Localisation Businesses). The survey was conducted between August 6, 2020 and September 3, 2020, eliciting 52 responses in total, and focused on business context, translation and localization customers, and market drivers. Key findings were compiled by Slator. Commentary on key findings was provided by the following language industry professionals based in India: The majority of survey respondents are from small LSPs, with revenues of less than USD 1m. Most have fewer than 10 employees. Some 20–25% are slightly larger with a headcount of 10–50 employees and are forecasting revenues of USD 1–8m in 2020. Sandeep Nulkar, President of Indian language industry association, CITLoB, shared his thoughts on the typical size of LSPs in India: “USD 1m converted into INR is a significant amount of money,” he said. “I would imagine that there will not be more than a few purely Indian and purely translation- and allied-service-centric LSPs doing that kind of revenue.” Many LSPs are based in Pune, Delhi, and Bangalore. Delhi is the top location for the clients of LSPs, 69% of which have clients based there. The top two business challenges for Indian LSPs right now are sales / growth and the availability of suitable linguistic resources. A little more than half are seeing a decline in rates, while most believe that the dominant pricing model for written translation services will be hybrid by 2022. CITLoB’s Nulkar offered some advice for LSPs looking to increase sales and business growth: “Prospecting and building sales funnels is an art, and one needs to hire professional sales people to do that. Most [small to mid-size] LSPs cannot afford it. If they form the majority of our respondents, then clearly they would say that sales and growth is a challenge; but then I would imagine that is from an inability-to-sell point of view rather than any fundamental issue with demand. My advice is just that. Leave selling to the experts. If you happen to be one and also run an LSP, fine. If not, accept it and hire the best you can within your budget. India has a huge workforce and a lot of hungry salespeople. You will find the right person if you are hungry enough.” Looking beyond 2020, many LSPs have a very positive outlook for 2021. Most report little or no negative impact from Covid-19, while around a fifth said the negative impact from Covid-19 had been significant. Regarding most LSPs not seeing an impact on business as a result of Covid-19, Nulkar commented that “all major players I know have been severely affected [by Covid].” He added that some LSPs may be “in denial” about the impact on their revenues. He offered another explanation: some LSPs may derive their business from “very few sources,” which were lucky enough not to be affected that much. Several survey questions focused on the use of and attitudes toward machine translation (MT) and language technology. MT integration is relatively not advanced among the LSPs represented in the survey. More than 20% have not started the process of integrating MT into their production workflow, and only a handful have fully-implemented and scalable MT capabilities. Most are early- to mid-integration, while nearly 20% have almost completed the journey. MT topped the list as the greatest external threat to Indian LSPs and the biggest industry megatrend. Compared to the responses in Slator’s survey of 50 globally-located LSPs, published in May 2020, those in Slator’s Indian LSP survey show less advanced implementation of MT: the number of respondents in the global LSP survey who said MT was fully implemented and scalable in their production workflow was 2.5 times that of the Indian LSP survey. The area of AI / automation / technology closely follows MT as the second greatest external threat to LSPs. Meanwhile, popular megatrends for Indian LSPs include improvements in translation productivity products (CAT), advancements in language technology other than MT, and improvement in translation management products (TMS). When it comes to translation productivity tools (CAT and TMS), more than three-quarters of Indian LSPs use third-party software. Notably, the majority only use third-party translation tools (with no proprietary technology). The LSPs said they used 18 different third-party tools overall. By far the most popular third-party tool is SDL Trados, used by nearly 40% of LSPs. The next most popular tools cited by respondents included memoQ and Memsource, followed by Rian (a translation software provider based in India), Smartcat, and Wordfast. Discussing the level of MT integration within Indian LSPs, CITLoB’s Nulkar said, “Again, it is a money issue. If your revenue is not significant enough, then your idea of MT is Google. Investing in MT is costly and especially in INR (Indian rupee) because generic engines won’t work. You would need language pair- and domain-specific ones so that would be a lot of money if an LSP offers even a few language pairs.” He added: “Whether [or not] it is a matter of time would depend on how MT companies are able to think ‘wholesale prices’ specific to India. India is a price sensitive market — even fancy cars sell at a fraction of their EU prices. Unless MT vendors are able to [offer] India pricing, they can forget about achieving scale. But if they do, I am sure they will make most of their money here.” Nulkar characterized the general attitude of Indian LSPs toward MT as being one of suspicion: “In their head it is equal to Google Translate and, therefore, equal to more work, although Google Translate is getting better even with some Indian languages. It was the same attitude when CAT tools arrived on the scene. CAT Tools [meant] ‘I will now not get paid for everything.’ It’s not a growth mindset.” When it comes to the widespread use of third-party translation technology among Indian LSPs, Nulkar, who is Founder, Chairman, and Managing Director of LSP BITS in addition to his role at CITLoB, said it was a question of cost, revenue, and prioritization. He cited his own company as an example: BITS built its own client relationship management (CRM) system and then used it for 10 years before deciding to switch to a third-party CRM. Nulkar said this was “because we wanted our IT team to build the Vernac platform (and earn us money), not service a CRM (and save us money); and because we were not rich enough as a company to do both.” (Vernac is a crowdsourced translation and localization solution.) He added, “For someone with even lesser revenue building their own tools, CRM or any other, it probably does not even cross their mind.” On the whole, interpreting is not big business for LSPs in India: nearly 90% derive less than 10% or none of their revenues from interpreting. Other than the core services of translation and localization, LSPs also offer a range of non-core services, such as subtitling, voiceover / dubbing, language training / teaching, and transcription, which emerged as the most popular ancillary services. The top ancillary services for LSPs broadly overlap with those that LSPs identify as offering the biggest growth opportunities: subtitling, AI-related services, voiceover / dubbing, MT-related services, and language training. The LSPs represented in the survey are mainly multi-sector, working across an average of 5.3 end-customer verticals. The top customer segments for LSPs are Technology, which is serviced by nearly 80% of all LSPs; Professional Services, which includes sectors such as edutech, e-learning, market research, and legal; and Engineering & Manufacturing. Sudeip Kummar, Founder at White Globe Pvt Ltd, said that “the Indian market is not as evolved and mature compared to Asia, Europe, and the Americas.” This means that, “currently, customers give more importance to language skills compared to industry specialization and all LSP’s work across multiple sectors,” he added, predicting that “specialization would happen as the market matures.” Kummar also explained the importance of India’s technology and IT sector to the economy, saying that “the IT Sector in India is renowned for its capabilities, has a multinational presence, [and makes] a large contribution to the national economy.” On the role of language services within the technology sector, he said, “Most of these organizations need language support for their MNC customers across the globe; from sales support to delivery and post-sales. The headquarters are located in India and there is easier access for Indian LSPs. Most of the LSPs in India have a significant [revenue] share from this segment.” Madhu Sundaramurthy, Managing Director – APAC for Summa Linguae Technologies, agreed with this assessment. She said, “India is a hub for IT companies like Infosys, Accenture, Wipro, and Cognizant. These companies provide IT software and services to customers across the globe and translation is a part of [it]. This is why most LSPs in India work with IT and technology companies.” Commenting on the multi-sector nature of most LSPs, Sundaramurthy said, “It is imperative that LSPs are ready to work in all domains, [because] technology companies work with different clients from different domains. Very few LSPs specialize in one domain like medical or pharma.” A reflection of the diversity of serviced verticals, LSPs were divided on which segment represents the biggest opportunity for their business; although Life Sciences, Professional Services, Technology, and Media & Gaming emerged as the marginal frontrunners. More than 50% of LSPs work across the Media, Finance, and Life Sciences segments; but less than 20% currently work in Gaming. Only a third of LSPs said they subcontract from larger LSPs, indicating that most work directly with end-customers or other types of middlemen. Sundaramurthy and Kummar both observed that subcontracting from larger LSPs is not very common in India. According to Sundaramurthy, “In India, very few LSPs subcontract to smaller LSPs,” but there are some LSPs that “only work with other LSPs in India and outside.” Kummar confirmed that most LSPs work directly with clients. “However, there are specific segments where LSPs also work with partners like marketing agencies, etc.” He said some LSPs subcontract from larger LSPs, but this is not particularly prevalent. LSP’s mainly work with customers based across India and internationally. A fourth work with customers throughout India but not international customers, while relatively few primarily work with customers local to the region where they are based. The biggest international market for Indian LSPs is Asia and then Europe, both of which are serviced by the majority of LSPs represented in the survey; 35% of LSPs work with customers based in North America. White Globe’s Kummar said that the proportion of LSPs working for customers in North America is low “primarily because of the lack of presence in international markets for most of the Indian LSPs.” However, he believes that, “as the market matures in India and consolidation happens, a few of the leading players will definitely have a global presence.” This will allow LSPs to increase the percentage of revenues and absolute amount they derive from these regions, Kummar said. Sundaramurthy also commented on these findings, saying that Indian LSPs’ relative lack of a customer base in North America is “connected with the IT and technology companies.” She explained: “Many international companies like Microsoft, Google, Adobe have offices in India and outsource work to Indian companies. So it may seem that we are working with only Asian clients but it is the APAC office of international clients.” Like Kumamar, Sundaramurthy also believes that market consolidation will lead to LSPs working with more North American customers directly. She said, “Direct projects with clients in the US are bound to increase with all the mergers and acquisitions between Indian and international LSPs.” More than two-thirds of LSPs pegged Hindi as the most important target language based on volumes requested. Out of India’s 22 official languages, 14 were not selected by any LSP as representing the biggest demand. More than half the LSPs said there was significant growth in demand for Indian languages from both domestic and international customers. Another third said there was a slight increase from Indian customers, while 29% said there was a slight increase from international customers. The biggest demand driver for LSPs is their clients’ desire to reach more end-customers or expand the business. As the second most populous country in the world after China, India is an attractive market for commerce and trade and, by localizing content into more Indian languages, companies can hope to reach a bigger audience and additional customers across the country. Sunil Kulkarni, CEO of Fidel Softech Pvt Ltd, explained some of the factors driving LSPs’ end-customer interest in India’s economy: “Within the huge population, there are multiple India’s with different economic criteria. So for companies, too, it becomes a sizable market for expansion with different price points and different products or services.” According to Kulkarni, India also serves as a testing ground and an export hub for other comparable markets: “The same market is also useful for companies to sample or try out new product development and then launch it in similar emerging markets” (e.g., Asean, Africa, or the Middle East). Kulkarni cited the Nissan plant in Chennai, where hardly 10% is manufactured for the Indian market, with the rest being exported to the Middle Eastern and African markets. “Hence India is not [only] being seen for its population, but also from a product development / R&D / logistics / labor perspective as well.” LSPs also identify content creation and streaming platforms as important factors in driving demand from international customers, more so than from domestic customers. The same is true of the government’s “Make in India” initiative, which is designed to attract foreign capital and encourage international companies to do business in India. On a sector basis, LSPs believe e-learning to be the most important driver of demand for their services. More than 90% of respondents rated e-learning a four or five (out of five) in terms of importance for their business. E-commerce is the second most important sector-driver of demand. On the lower end of the scale are the drivers of domestic fintech and regulatory changes, which were rated four or five by 60% of respondents. The sector drivers of big tech firms, domestic tech firms, OTT / streaming, and smartphone use emerged somewhere in the middle. Several questions focused on the experiences of LSPs within specific sectors, and LSPs were asked to identify emerging trends in the sectors of e-commerce, OTT, and e-learning. E-commerce: Many LSPs said that demand for e-commerce localization is increasing. LSPs also identify the following trends: Fidel Tech’s Kulkarni said that e-commerce is indeed one of the biggest growth areas for LSPs: “With the proliferation of smartphones and the cheapening of Internet data packages, rural and semi-urban area consumers also demand or expect equal services from their urban counterparts.” Given Indian’s multilingualism, delivering e-commerce for India means more than a hefty translation demand. It also requires companies to put in place a localized e-commerce experience. Kulkarni explained: “E-commerce for such a diverse set of consumers means localizing of UI in local languages as well as usage of voice technologies for search, SEO-ing for a range of words as well as allowing searches in local languages on the website. This also means that the logistics and last-mile applications are also localized — as the local delivery boy needs to identify the customer and the correct address.” OTT: While more than half of LSPs service customers in the Media segment, just over a quarter of LSPs currently work with streaming services or OTT providers; 14% said Media is the top growth segment for their business. Kulkarni said that some OTT platforms outsource content creation services to third-party companies, meaning that LSPs may receive work destined for OTT platforms via a middleman. He said, “OTT streaming customers are offloading some parts of content creation (e.g., voiceover, subtitling, mobile app creation, or hosting infrastructure) to some media firms. In some cases, media companies are leasing their studios to OTT players and, hence, some contracts are seen through direct OTT players and some through media firms. The media customers include advertising agencies, voice studio firms, video and mobile technology firms.” Subtitling, dubbing, and voiceover are all important non-core services for LSPs, but their application is not strictly limited to OTT or streaming and Media customers. E-learning, the biggest sectoral demand driver for localization in India, also calls for audio-visual localization services. E-learning: Most LSPs said that demand for e-learning localization is increasing. LSPs also made several specific observations, crediting growth to a variety of factors, such as Covid-19, online language learning, professional training, and higher education. Other trends identified in e-learning include its increasing prevalence in remote parts of India, the demand for more vernacular languages, and bite-sized learning. According to Kulkarni, a specific niche of e-learning, edutech, is surging in India. “Edutech is seeing a huge growth, especially in these pandemic times.” Explaining the opportunity for LSPs, he said edutech “involves content development [and the] convergence of voice and video technologies as well. LSPs can forge partnerships or develop in-house capabilities in any or few of these areas to capture the market.” Kulkarni believes that there is a long-term opportunity for LSPs in this niche, resulting from ongoing development requirements as the amount of content and the size of the market increase. Edutech is not only concerned with childhood education; it also includes adult education and training. For example, Kulkarni said, “under the JanDhan scheme, the Indian government [enabled] marginalized people in society to open bank accounts. But then they needed to be taught about bank transfer or ATM usage and so on. Here, local language UI-based educational content for digital payments, fintech, or insurance were sought and huge opportunities were created.” Respondents are located throughout India, with the popular locations being Pune (18), Delhi (10), Bangalore (8), Chennai (5), and Mumbai (4). Two survey respondents said their businesses are located outside India, in Europe. The vast majority of respondents hold a Master’s degree or equivalent as their highest academic qualification (77%), while some hold an undergraduate degree or equivalent (17%). Some 44% of respondents hold at least one academic qualification related to translation, interpreting or languages, while 56% have no language-related academic qualifications. Respondents represent LSPs of various sizes. Around 60% work in an LSP with less than 10 employees; 21% work in an LSP with 10–50 employees; 12% with 50–200 employees; and 8% with more than 200 employees. Of the 33 respondents who answered the question relating to company revenues, most work for a Boutique LSP, with annual revenues of USD 8m or less. More than 60% expect to generate less than USD 1m in 2020 — a similar proportion to those who said they work in LSPs with less than 10 employees — and another quarter expect 2020 revenues of USD 1–8m. Respondents’ estimates on the size of their LSP’s addressable market varied greatly, although the majority of respondents placed their addressable market size at less than USD 100m (82%). More than half of all respondents said that unit rates were declining, while 38% said they were stable. Only 10% said they are seeing an increase in unit rates. More than half of all respondents expect the dominant pricing model for written translation services to be “Hybrid” by 2022, while more than a third believe the “Per word” model will remain dominant. Respondents identified accelerating progress in machine translation (MT) as the most important megatrend in the global language industry in 2020/1 (42%). The next most important megatrends were identified as improving translation productivity products (15%) and accelerating progress in language technologies other than MT (12%). Interpretation is not a big business for survey respondents; nearly 90% said they do not do any interpreting or generate less than 10% of their business from interpretation. Respondents were positive overall in terms of their outlook for the language industry for 2021. Nearly 90% felt very or somewhat positively about the future. A further 13% were neutral, but none were somewhat or very negative. Most respondents reported no or minimal impact from Covid-19. More than 80% said the impact had been slight (in either direction) or nonexistent; 19% reported a significant negative impact, while no respondents reported a very positive impact. Most respondents said their gross margin target for 2020 is 10–39% (61%). Slator grouped respondents’ answers to this free-text question into categories to capture the nature of the threats identified. The most significant type of external threat for respondents was machine translation (MT), closely followed by a related category, that of AI / automation / technology. Within the category of “Competitive landscape,” which emerged as the third biggest type of threat identified, respondents cited concerns such as undercutting of prices by competitors and new market entrants, the role of big tech in the language industry, and independent middlemen. Price pressure and the availability of suitable resources were also identified as threats, although by significantly fewer respondents than the top three threats. There were also a number of “other” threats identified by respondents that only received one mention (not displayed above). These included Covid-19, the increasing prevalence of English as a global language, and politically-motivated trade restrictions. A handful of respondents did not identify any external threats, while some provided multiple responses. Respondents were invited to list their current top three business challenges. Slator then grouped respondents’ answers to this free-text question into categories to capture the nature of the challenges identified. The biggest challenge for respondents was related to sales and business growth, closely followed by the availability of suitable linguistic resources. Price pressure, MT, and the challenge of educating or managing clients emerged as the third, fourth, and fifth biggest business challenges, respectively. Outside of these five categories, respondents also said that people / talent management (17%), the competitive landscape (15%), and technology (12%) currently represented challenges to their business. Covid-19, deadlines, marketing, and remote working were each mentioned as challenges by 6% of respondents. Nearly 60% of respondents exclusively use third-party TMS and CAT tools, while 17% use a mix of proprietary and third-party tools. A small percentage (7%) said that they only use proprietary language tools, while a further 17% said they use no TMS or CAT tools or did not provide a response. Respondents were invited to specify which translation tools they use. By far, the most popular third-party tool for respondents is SDL Trados, used by 38% of all respondents. Other SDL products were used by 6% of respondents. The next most popular tools include memoQ (12%) and Memsource (10%). Rian, an India-based translation software, Smartcat, and Wordfast are used by 8%, 6%, and 6% of respondents, respectively. Meanwhile, 6% said they use non-translation specific tools (e.g., Odoo, ExaVault). Relatively few respondents (5.8%) have completed the process of fully integrating machine translation (MT) into their production workflow. A greater number, more than 20%, have not started MT integration. A similar number (19.2%) have mostly completed the process of implementation (7/10 – 9/10), while 38% are mid-implementation (4/10 – 6/10). A further 15% are in the early stages of implementation (2/10 – 3/10). Over time, one would expect the bar chart to trend to the right as companies currently reporting lower levels of integration progress toward higher levels of MT integration. Respondents provide a combined total of 42 ancillary services to their customers (other than translation, localization, and interpreting), and an average of 2.2 each. The top three ancillary services for respondents to provide are subtitling, voiceover, and language training or teaching, in that order. Transcription, dubbing, software engineering, formatting / DTP, and content writing are also relatively popular services among respondents, while other services listed are provided by fewer respondents. The services mentioned by a single respondent are not listed in the chart above. These are (listed alphabetically) as follows: Audio recording, CAD-CAM services, CAT tool services, Chatbots, Content development, Corpus alignment, Corpus building, Creative writing, Data annotation, Data creation (AI), Data entry, Digital e-publishing, Legal services, Linguistic services, Managed Services, Mixing, MTPE, Multilingual data services, Multilingual staffing, NLP, Post-production services, Research analysis, Reverse proxy solution, Reviewing, and Translation training. Important to note in particular is that MTPE may well be offered by more than one respondent’s LSP, but others may consider it core rather than ancillary to translation services. The same is true of a number of other services, such as CAT tool services, editing, and proofreading. Respondents service 5.3 end-customer verticals on average. The most popular end-customer vertical among respondents is Technology (79%), followed closely by Professional Services (73%), with Engineering & Manufacturing customers a more distant third. More than half the respondents work for customers across Life Sciences, Finance, and Media. The least common verticals for respondents are Aerospace & Defense (13%), followed by Gaming (19%). The picture is mixed as to which end-customer vertical represents the biggest growth opportunity for respondents’ businesses, with none emerging as a convincing frontrunner. Life Sciences, Professional Services, and Technology received the most responses, with 18%, 16%, and 16%, respectively. Gaming and Media were both identified as the biggest growth opportunity by more than 10% of respondents. Most remaining end-customer segments received relatively few responses, while no respondents identified Aerospace & Defense as the biggest growth opportunity for their business. Within the broader umbrella of Professional Services, a number of respondents specifically identified e-learning and / or education as important growth segments. Most respondents said their customers are based across India and internationally (62%). A fourth (25%) of respondents have customers located nationwide (but not internationally), while a smaller percentage primarily serve customers based locally in the region where they are located. The top 10 Indian locations for respondents’ customers are Delhi, Karnataka, Andhra Pradesh, Gujarat, Chandigarh, Maharashtra, Haryana, Goa, Assam, and Bihar. Nearly 70% of respondents have customers located in Delhi and a little over half have customers located in Karnataka. Beyond these locations, respondents’ customers were dispersed fairly evenly across the country. More than 70% of respondents do business elsewhere in Asia, while more than half operate in Europe. A little more than a third of respondents have business operations in North America; 13% said they do not operate internationally in any of these three continents. Of the 45 people who responded to this question, two-thirds said they do not subcontract work from other LSPs, while a third said they do. Nearly 60% of respondents said that Hindi is the most highly requested target language, while a far smaller percentage identified Marathi (17%), Tamil (13%), or Kannada (3.9%) as the target language that represents the biggest volumes for their business. One respondent per language said that Assamese, Bengali, Malayalam, and Telugu provided the biggest volumes, while the remaining 14 official languages of India were not selected by any respondent. More than 50% of respondents said that there is significant growth in demand from both international and domestic customers for Indian languages in 2020. A similar percentage said there is slight growth for Indian languages from international customers as from domestic customers (29% vs. 33%); 13% of respondents said they are observing flat (no) growth for Indian languages from international customers, while 19% said demand was flat from domestic customers. Opinions were divided among respondents as to whether regulatory changes constitute a driver of demand for translation, localization, and interpreting in India; 44% of respondents are unsure, while 42% said changing regulation is a demand driver. Respondents said that the biggest demand driver for Indian languages from both domestic and international clients is their desire to reach more end-customers or expand their business. Respondents also identified specific end-customer segments or content types (such as Life Sciences, Edutech, Marketing, and Media) as an important driver of demand from domestic and international customers. Responses reveal that content creation or platforms (e.g., streaming platforms) is a bigger factor in driving demand from international customers than domestic customers. This is also the case for the government’s “Make in India” initiative, which is designed to encourage international companies to do business in India. In total, respondents identified 27 ancillary services as growing in demand. The top five growth services among respondents are subtitling, AI-related services, voiceover / dubbing, MT-related services, and language training. When asked what customer trends they observe in e-commerce localization, many respondents (37%) answered that demand was increasing. Others identified specific trends in e-commerce; for example, one respondent noted that e-commerce companies in India have begun bringing localization activities in-house. Two respondents observed that there was demand for more languages, while three said that there was a trend toward using more machine translation (either MTPE / PEMT or MT) for e-commerce localization. Respondents were asked whether they work for any OTT (over-the-top) providers such as overseas streaming services, Netflix and Amazon, or local streaming services including Hotstar and Viu. Most respondents said they do not currently work with this type of customer (73%), while just over a quarter said they do (27%). When asked what customer trends they observe in e-learning, most (65%) responded or implied through their answers that demand was increasing in this segment. Some qualified their responses. For example — Other trends observed in e-learning included more vernacular languages and bite-sized learning. Respondents were asked to rate, out of 5, the importance for their business of specific demand drivers, 1 being “not at all important” and 5 being “very important.” The most important driver for respondents is e-learning, which more than 92% of people rated 4 or 5. The next most important driver is e-commerce, which received 84% 4s and 5s. On the lower end of the scale were Domestic fintech and Regulatory changes, which were rated 4 or 5 on the scale by 60% of respondents. Note: Survey participants did not necessarily respond to all questions.
Slator 2020印度语言服务提供商 (LSP) 调查是由Slator与印度语言行业协会CITLoB (口译,翻译和本地化业务联合会) 合作进行的。 该调查是在2020年8月6日和2020年9月3日之间进行的,总共获得了52份答复,重点是业务背景,翻译和本地化客户以及市场驱动因素。 关键发现由Slator汇编。以下印度语言行业专业人士提供了对主要发现的评论: 大多数受访者来自小型lsp,收入不到100万美元。大多数员工少于10名。大约20-25% 人的人数稍大,员工人数为10-50人,预计2020年收入为1-800万美元。 CITLoB印度语言行业协会主席Sandeep Nulkar分享了他对印度lsp典型规模的看法: “将100万美元转换为INR是一笔不小的数目,” 他说。“我可以想象,不会有超过几个纯粹的印度和纯粹的翻译和盟军服务为中心的lsp从事这种收入。” 许多lsp都位于浦那,德里和班加罗尔。德里是lsp客户的首选位置,其中69% 有客户在那里。 目前,印度lsp面临的两个最大业务挑战是销售/增长以及合适的语言资源的可用性。略多于一半的人看到费率下降,而大多数人认为书面翻译服务的主要定价模式将是混合2022年。 CITLoB的Nulkar为希望增加销售和业务增长的lsp提供了一些建议: “勘探和建立销售漏斗是一门艺术,需要聘请专业的销售人员来做到这一点。大多数 [中小型] lsp负担不起。如果他们占我们受访者的大多数,那么显然他们会说销售和增长是一个挑战; 但我可以想象这是从无法销售的角度出发,而不是从需求的任何根本问题出发。我的建议就是这样。把卖给专家。如果您碰巧是一个,并且还运行LSP,那就可以了。如果没有,接受它,并在预算范围内聘请最好的。印度拥有庞大的劳动力和大量饥饿的销售人员。如果你足够饿,你会找到合适的人。“ 除了2020之外,许多lsp对2021有着非常积极的前景。大多数人报告新型冠状病毒肺炎的负面影响很小或没有,而大约五分之一的人表示新型冠状病毒肺炎的负面影响很大。 关于大多数lsp没有看到Covid-19对业务的影响,Nulkar评论说: “我认识的所有主要参与者都受到 [Covid] 的严重影响。” 他补充说,一些lsp可能 “否认” 对其收入的影响。他提供了另一种解释: 一些lsp可能来自 “很少的来源”,这些来源很幸运,没有受到太大影响。 几个调查问题集中在对机器翻译 (MT) 和语言技术的使用和态度上。 在调查中代表的lsp中,MT集成相对不先进。超过20% 尚未开始将MT集成到其生产工作流程中,并且只有少数具有完全实现和可扩展的MT功能。大多数是早期到中期的整合,而近20% 几乎已经完成了旅程。MT是印度lsp面临的最大外部威胁和最大的行业大趋势,位居榜首。 与Slator在2020年5月发布的对50个全球LSP的调查中的回答相比,Slator的印度LSP调查中的回答显示MT的实施不那么先进: 全球LSP调查中表示MT在其生产工作流程中已完全实施且可扩展的受访者人数是印度LSP调查的2.5倍。 人工智能/自动化/技术领域紧随MT之后,是lsp面临的第二大外部威胁。同时,印度lsp的流行大趋势包括翻译生产力产品 (CAT) 的改进,MT以外的语言技术的进步以及翻译管理产品 (TMS) 的改进。 当谈到翻译生产力工具 (CAT和TMS) 时,超过四分之三的印度lsp使用第三方软件。值得注意的是,大多数人只使用第三方翻译工具 (没有专有技术)。 Lsp表示,他们总体上使用了18种不同的第三方工具。到目前为止,最受欢迎的第三方工具是SDL Trados,近40% 的lsp使用它。受访者引用的下一个最受欢迎的工具包括memoQ和Memsource,其次是Rian (印度的翻译软件提供商),Smartcat和Wordfast。 在谈到印度lsp内部的MT整合水平时,CITLoB的Nulkar说: “同样,这是一个资金问题。如果你的收入不够可观,那么你对MT的想法就是谷歌。投资MT的成本很高,尤其是INR (印度卢比),因为通用发动机无法正常工作。你需要语言对和特定于领域的语言,所以如果一个LSP提供几个语言对,那将是一大笔钱。“ 他补充说: “这是否是时间问题,将取决于MT公司如何思考印度特有的 '批发价'。印度是一个价格敏感的市场-即使是高档汽车的售价也只是欧盟价格的一小部分。除非MT供应商能够 [提供] 印度定价,否则他们可能会忘记实现规模。但如果他们这样做了,我相信他们会在这里赚大部分钱。“ Nulkar将印度lsp对MT的普遍态度描述为怀疑之一: “在他们的脑海中,它等于Google翻译,因此等于更多的工作,尽管Google翻译即使使用某些印度语言也在变得更好。当CAT工具到达现场时,态度也是如此。猫工具 [的意思是] '我现在不会得到所有的报酬。' 这不是一种增长心态。“ 当谈到第三方翻译技术在印度LSP中的广泛使用时,除了在CITLoB担任职务外,LSP BITS的创始人、董事长兼董事总经理Nulkar表示,这是一个成本、收入和优先级的问题。他以自己的公司为例: BITS建立了自己的客户关系管理 (CRM) 系统,然后使用了10年,然后才决定改用第三方CRM。 Nulkar说,这是 “因为我们希望我们的IT团队建立Vernac平台 (并为我们赚钱),而不是为CRM服务 (并为我们省钱); 并且因为我们作为一家公司还不够富有,无法做到这两者。” (Vernac是一个众包翻译和本地化解决方案。)他补充说: “对于那些拥有更少收入来构建自己的工具 (CRM或任何其他工具) 的人来说,他们甚至可能不会想到。” 总体而言,口译对于印度的lsp来说并不是一件大事: 近90% 的人从口译中获得的收入不到10% 或没有。 除了翻译和本地化的核心服务外,lsp还提供一系列非核心服务,例如字幕,配音/配音,语言培训/教学和转录,这些服务已成为最受欢迎的辅助服务。 Lsp的顶级辅助服务与lsp认为提供最大增长机会的服务大致重叠: 字幕,AI相关服务,配音/配音,MT相关服务和语言培训。 调查中代表的lsp主要是多部门的,平均适用于5.3个最终客户垂直行业。Lsp的顶级客户群是技术,由几乎80% 的lsp提供服务; 专业服务,包括edutech,电子学习,市场研究和法律等领域; 以及工程与制造。 White Globe Pvt Ltd的创始人Sudeip Kummar说: “与亚洲,欧洲和美洲相比,印度市场还没有发展成熟。” 这意味着,“目前,与行业专业化和LSP跨多个部门的所有工作相比,客户对语言技能的重视程度更高。” 他补充说,并预测 “随着市场的成熟,专业化将会发生。” Kummar还解释了印度技术和IT部门对经济的重要性,他说: “印度的IT部门以其能力而闻名,拥有跨国公司,并为国民经济做出了巨大贡献。” 关于语言服务在技术领域中的作用,他说: “这些组织中的大多数都需要为全球范围内的跨国公司客户提供语言支持; 从销售支持到交付和售后。总部位于印度,印度lsp更容易访问。印度的大多数lsp在这一细分市场都有很大的 [收入] 份额。“ Summa Linguae Technologies亚太地区董事总经理Madhu Sundaramurthy同意这一评估。她说: “印度是Infosys、埃森哲、Wipro和Cognizant等IT公司的枢纽。这些公司为全球客户提供IT软件和服务,翻译是 [it] 的一部分。这就是印度大多数lsp与IT和技术公司合作的原因。“ Sundaramurthy在评论大多数lsp的多部门性质时说: “lsp必须准备好在所有领域中工作,因为技术公司与来自不同领域的不同客户合作。很少有lsp专攻医疗或制药等领域。“ 反映了服务垂直行业的多样性,lsp在哪个细分市场代表其业务的最大机会上存在分歧; 尽管生命科学,专业服务,技术以及媒体和游戏成为边缘领先者。 超过50% 的lsp在媒体,金融和生命科学领域开展工作; 但目前从事游戏工作的不到20%。 只有三分之一的lsp表示他们从较大的lsp分包,这表明大多数lsp直接与最终客户或其他类型的中间商合作。 Sundaramurthy和Kummar都观察到,从较大的lsp分包在印度并不常见。根据Sundaramurthy的说法,“在印度,很少有lsp分包给较小的lsp,” 但是有些lsp “仅与印度内外的其他lsp合作”。 Kummar确认大多数lsp直接与客户合作。“但是,在某些特定领域,lsp也与营销机构等合作伙伴合作。” 他说,一些lsp从较大的lsp分包,但这并不是特别普遍。 LSP主要与印度和国际上的客户合作。第四个是与印度各地的客户合作,而不是与国际客户合作,而相对较少的主要是与他们所在地区的本地客户合作。 印度lsp的最大国际市场是亚洲,然后是欧洲,这两个市场都由调查中代表的大多数lsp提供服务; Lsp的35% 与北美客户合作。 White Globe的Kummar表示,为北美客户服务的lsp比例较低,“主要是因为大多数印度lsp在国际市场上缺乏业务”。 然而,他认为,“随着印度市场的成熟和整合的发生,一些领先的参与者肯定会在全球范围内占有一席之地。” Kummar说,这将使lsp能够增加他们从这些地区获得的收入和绝对数量的百分比。 Sundaramurthy还评论了这些发现,称印度lsp在北美相对缺乏客户群是 “与IT和技术公司有关”。她解释说: “许多国际公司,如微软,谷歌,Adobe在印度设有办事处,并将工作外包给印度公司。因此,我们似乎只与亚洲客户合作,但它是亚太地区国际客户办事处。“ 与Kumamar一样,Sundaramurthy也认为,市场整合将导致lsp直接与更多北美客户合作。她说: “随着印度和国际lsp之间的所有并购,与美国客户的直接项目必将增加。” 超过3分之2的lsp根据所请求的卷将印地语作为最重要的目标语言。在印度的22种官方语言中,没有任何LSP选择14种代表最大的需求。 超过一半的lsp表示,国内外客户对印度语言的需求都有显着增长。另有三分之一的人表示,印度客户略有增加,而29% 则表示,国际客户略有增加。 Lsp的最大需求驱动因素是他们的客户希望接触更多的最终客户或扩大业务。作为仅次于中国的世界第二大人口大国,印度是一个有吸引力的商业和贸易市场,通过将内容本地化为更多的印度语言,公司可以希望在全国范围内吸引更多的受众和更多的客户。 Fidel Softech Pvt Ltd首席执行官Sunil Kulkarni解释了一些推动lsp最终客户对印度经济感兴趣的因素: “在庞大的人口中,有多个印度具有不同的经济标准。因此,对于公司来说,它也成为一个相当大的市场,可以通过不同的价格点和不同的产品或服务进行扩张。“ 根据Kulkarni的说法,印度也是其他可比市场的试验场和出口枢纽: “相同的市场对于公司采样或尝试新产品开发,然后在类似的新兴市场推出也很有用” (例如,东盟,非洲,或中东)。 Kulkarni引用了日产在钦奈的工厂,那里几乎没有为印度市场生产10%,其余的则出口到中东和非洲市场。“因此,印度不仅 [] 被视为其人口,而且从产品开发/研发/物流/劳动力的角度来看也是如此。” Lsp还将内容创建和流媒体平台视为推动国际客户需求的重要因素,而不是国内客户。政府的 “印度制造” 计划也是如此,该计划旨在吸引外国资本并鼓励国际公司在印度开展业务。 就行业而言,lsp认为电子学习是其服务需求的最重要驱动力。就其业务的重要性而言,超过90% 的受访者将电子学习评为4或5 (满分5)。电子商务是第二大最重要的部门-需求驱动力。 规模较低的是国内金融科技和监管变化的驱动力,60% 的受访者将其评为四到五。大型科技公司、国内科技公司、OTT/流媒体和智能手机使用的行业驱动力出现在中间。 几个问题集中在特定部门的lsp经验上,并要求lsp确定电子商务,OTT和电子学习领域的新兴趋势。 电子商务: 许多lsp表示,对电子商务本地化的需求正在增加。Lsp还确定了以下趋势: Fidel Tech的Kulkarni表示,电子商务确实是lsp最大的增长领域之一: “随着智能手机的普及和互联网数据包的廉价化,农村和半城市地区的消费者也要求或期望城市同行提供平等的服务。” 鉴于印度人的多种语言,为印度提供电子商务不仅意味着巨大的翻译需求。它还要求公司建立本地化的电子商务体验。Kulkarni解释说: “针对如此多样化的消费者的电子商务意味着本地化本地语言的UI,以及使用语音技术进行搜索,搜索一系列单词的SEO以及允许在网站上以本地语言进行搜索。这也意味着物流和最后一英里的应用程序也本地化了-因为本地的送货员需要识别客户和正确的地址。” OTT: 虽然媒体领域一半以上的lsp服务客户,但目前只有四分之一以上的lsp与流媒体服务或OTT提供商合作; 14% 表示,Media是其业务增长最快的部分。 Kulkarni说,某些OTT平台将内容创建服务外包给第三方公司,这意味着lsp可能会通过中间人接收发往OTT平台的工作。他说: “OTT流媒体客户正在将内容创建的某些部分 (例如,画外音,字幕,移动应用程序创建或托管基础架构) 卸载给某些媒体公司。在某些情况下,媒体公司将其工作室出租给OTT播放器,因此,一些合同是通过直接OTT播放器获得的,一些合同是通过媒体公司获得的。媒体客户包括广告代理商,语音工作室公司,视频和移动技术公司。” 字幕,配音和配音都是lsp重要的非核心服务,但其应用并不严格限于OTT或流媒体和媒体客户。电子学习是印度本地化的最大部门需求驱动力,它也要求提供视听本地化服务。 电子学习: 大多数lsp表示,对电子学习本地化的需求正在增加。LSPs还进行了一些具体的观察,将增长归功于多种因素,例如Covid-19,在线语言学习,专业培训和高等教育。电子学习中发现的其他趋势包括其在印度偏远地区的普及,对更多白话语言的需求以及一口式学习。 根据Kulkarni的说法,电子学习的一个特定利基市场edutech在印度正在蓬勃发展。“Edutech正在看到巨大的增长,特别是在这个大流行时期。” 他在解释lsp的机会时说,edutech “涉及内容开发 (以及) 语音和视频技术的融合。Lsp可以在这些领域中的任何一个或几个领域建立合作伙伴关系或开发内部功能,以占领市场。” Kulkarni认为,随着内容量和市场规模的增加,持续的开发需求导致lsp在这一领域存在长期机会。 Edutech不仅关注儿童教育; 它还包括成人教育和培训。例如,库尔卡尼说: “根据JanDhan计划,印度政府 [使] 社会中的边缘化人群能够开设银行账户。但是随后需要向他们学习有关银行转帐或ATM使用情况等的知识。在这里,寻求基于本地语言UI的数字支付、金融科技或保险教育内容,并创造了巨大的机会。“ 受访者遍布印度各地,最受欢迎的地点是浦那 (18),德里 (10),班加罗尔 (8),钦奈 (5) 和孟买 (4)。两名受访者表示,他们的业务位于印度以外的欧洲。 绝大多数受访者拥有硕士学位或同等学历 (77%),而有些则拥有本科学位或同等学历 (17%)。 一些44% 的受访者至少持有一种与翻译、口译或语言相关的学历,而56% 没有语言相关的学历。 受访者代表各种规模的lsp。大约60% 人在员工少于10人的LSP中工作; 21% 在拥有10-50名员工的LSP中工作; 拥有50-200名员工的12%; 和拥有200多名员工的8%。 在回答与公司收入有关的问题的33位受访者中,大多数为精品LSP工作,年收入在800万美元或以下。超过60% 人预计将产生不到100万美元的2020年-与那些说他们在lsp中工作的员工人数少于10人的比例相似-另一个季度预计2020收入为1-800万美元。 受访者对LSP可寻址市场规模的估计差异很大,尽管大多数受访者认为其可寻址市场规模不到100万美元 (82%)。 超过一半的受访者表示单位利率正在下降,而38% 则表示稳定。只有10% 表示,他们看到单位费率增加。 超过一半的受访者预计书面翻译服务的主导定价模式将是 “混合” 2022年,而超过三分之一的受访者认为 “每字” 模式仍将占主导地位。 受访者认为,机器翻译 (MT) 的加速发展是2020/1年全球语言行业中最重要的大趋势 (42%)。接下来最重要的大趋势被确定为提高翻译生产率产品 (15%) 和加速除MT以外的语言技术的进步 (12%)。 对于受访者来说,口译并不是一件大事; 近90% 人表示,他们不做任何口译,也不会从口译中产生少于10% 的业务。 受访者对2021年语言行业的前景总体上是积极的。几乎90% 对未来感到非常积极。另一个13% 是中立的,但没有一个是多少或非常消极的。 大多数受访者报告新型冠状病毒肺炎没有影响或影响很小。超过80% 人表示,影响很小 (无论是哪个方向) 或不存在; 19% 报告了重大的负面影响,而没有受访者报告了非常积极的影响。大多数受访者表示,他们2020的毛利率目标是10-39% (61%)。 Slator将受访者对此自由文本问题的答案分为几类,以捕获所识别威胁的性质。对受访者来说,最重要的外部威胁类型是机器翻译 (MT),紧随其后的是相关类别,即人工智能/自动化/技术。 在 “竞争格局” 这一类别中,受访者提到了一些担忧,例如竞争对手和新市场进入者对价格的降低,大型技术在语言行业中的作用以及独立中间商。 价格压力和适当资源的可用性也被确定为威胁,尽管受访者比前三个威胁要少得多。 受访者还发现了一些 “其他” 威胁,只提到了一个 (上面没有显示)。其中包括新型冠状病毒肺炎,英语作为一种全球语言的日益普及,以及出于政治动机的贸易限制。少数答复者没有发现任何外部威胁,而一些答复者提供了多种答复。 邀请受访者列出他们目前面临的三大业务挑战。然后,Slator将受访者对此自由文本问题的答案分为几类,以捕获所确定挑战的性质。受访者面临的最大挑战是与销售和业务增长有关,紧随其后的是合适的语言资源的可用性。 价格压力,MT和教育或管理客户的挑战分别成为第三,第四和第五大业务挑战。 在这五个类别之外,受访者还表示,人员/人才管理 (17%),竞争格局 (15%) 和技术 (12%) 目前对他们的业务构成了挑战。新型冠状病毒肺炎、截止日期、营销和远程工作都被6% 受访者视为挑战。 近60% 的受访者专门使用第三方TMS和CAT工具,而17% 则混合使用专有和第三方工具。一小部分 (7%) 表示他们仅使用专有语言工具,而另一17% 则表示他们不使用TMS或CAT工具或未提供响应。 邀请受访者具体说明他们使用了哪些翻译工具。到目前为止,最受受访者欢迎的第三方工具是SDL Trados,38% 所有受访者都使用该工具。6% 的受访者使用了其他SDL产品。接下来最受欢迎的工具包括memoQ (12%) 和Memsource (10%)。Rian是一种基于印度的翻译软件,Smartcat和Wordfast分别被8%,6% 和6% 的受访者使用。同时,6% 说他们使用非翻译专用工具 (例如,Odoo,ExaVault)。 相对较少的受访者 (5.8%) 已经完成了将机器翻译 (MT) 完全集成到其生产工作流程中的过程。超过20% 的数量尚未开始MT集成。类似的数量 (19.2%) 大部分完成了实施过程 (7/10-9/10),而38% 是中期实施过程 (4/10-6/10)。另一15% 处于实施的早期阶段 (2/10-3/10)。 随着时间的流逝,人们会期望条形图会向右发展,因为目前报告较低的集成水平的公司会朝着较高水平的MT集成发展。 受访者总共向客户提供42项辅助服务 (翻译,本地化和口译除外),平均每项2.2项。 受访者提供的前三大辅助服务依次是字幕,配音和语言培训或教学。转录,配音,软件工程,格式化/DTP和内容编写也是受访者中相对受欢迎的服务,而列出的其他服务则由较少的受访者提供。 上表中未列出单个受访者提到的服务。这些 (按字母顺序列出) 如下: 录音,CAD-CAM服务,CAT工具服务,聊天机器人,内容开发,语料库对齐,语料库构建,创意写作,数据注释,数据创建 (AI),数据输入,数字电子出版,法律服务,语言服务,托管服务,混合,MTPE,多语言数据服务,多语言人员配备,NLP,后期制作服务,研究分析,反向代理解决方案,审查和翻译培训。 特别需要注意的是,MTPE很可能由多个受访者的LSP提供,但其他人可能认为它是核心的,而不是翻译服务的辅助。许多其他服务也是如此,例如CAT工具服务,编辑和校对。 受访者平均服务5.3最终客户垂直市场。受访者中最受欢迎的最终客户垂直方向是技术 (79%),紧随其后的是专业服务 (73%),工程和制造客户排在第三位。 超过一半的受访者为生命科学、金融和媒体领域的客户工作。受访者最不常见的垂直领域是航空航天和国防 (13%),其次是游戏 (19%)。 对于哪个最终客户垂直行业代表受访者业务的最大增长机会,情况喜忧参半,没有一个成为令人信服的领跑者。生命科学、专业服务和技术获得最多的响应,分别为18% 、16% 和16%。 超过10% 的受访者认为游戏和媒体都是最大的增长机会。大多数剩余的最终客户群收到的回复相对较少,而没有受访者认为航空航天与国防是其业务最大的增长机会。 在更广泛的专业服务范围内,许多受访者特别将电子学习和/或教育确定为重要的增长部分。 大多数受访者表示,他们的客户遍布印度和国际 (62%)。四分之一 (25%) 的受访者的客户位于全国范围内 (但不在国际范围内),而较小的比例主要为他们所在地区的本地客户提供服务。 受访者客户的印度十大地点是德里,卡纳塔克邦,安得拉邦,古吉拉特邦,昌迪加尔,马哈拉施特拉邦,哈里亚纳邦,果阿,阿萨姆邦和比哈尔邦。 近70% 的受访者的客户位于德里,一半以上的客户位于卡纳塔克邦。除这些地点外,受访者的客户在全国范围内分布相当均匀。 超过70% 的受访者在亚洲其他地方开展业务,而超过一半的受访者在欧洲开展业务。略高于三分之一的受访者在北美开展业务; 13% 说,他们在这三大洲中的任何一个都没有开展国际业务。 在回答这个问题的45人中,有3分之2人说他们不从其他lsp分包工作,而三分之一的人说他们这样做。 近60% 的受访者表示,印地语是要求最高的目标语言,而较小的比例将马拉地语 (17%),泰米尔语 (13%) 或卡纳达语 (3.9%) 确定为代表其业务最大数量的目标语言。 每个语言的一位受访者说,阿萨姆语,孟加拉语,马拉雅拉姆语和泰卢固语提供的数量最多,而印度其余14种官方语言没有被任何受访者选择。 超过50% 的受访者表示,国际和国内客户对印度语言2020年的需求都有显着增长。类似的百分比表示,来自国际客户的印度语言与来自国内客户的印度语言略有增长 (29% 比33%); 13% 的受访者表示,他们观察到来自国际客户的印度语言增长持平 (无),而19% 表示来自国内客户的需求持平。 对于监管变化是否构成印度翻译,本地化和口译需求的驱动力,受访者之间存在分歧; 44% 的受访者不确定,而42% 的受访者表示,法规的变化是需求的驱动力。 受访者表示,国内和国际客户对印度语言的最大需求驱动因素是他们希望接触更多的最终客户或扩大业务。受访者还确定了特定的最终客户群或内容类型 (例如生命科学,教育技术,市场营销和媒体) 是国内外客户需求的重要驱动力。 回应显示,与国内客户相比,内容创建或平台 (例如流媒体平台) 是推动国际客户需求的更大因素。政府的 “印度制造” 计划也是如此,该计划旨在鼓励国际公司在印度开展业务。 总体而言,受访者认为27项辅助服务的需求正在增长。受访者中排名前五的增长服务是字幕,AI相关服务,配音/配音,MT相关服务和语言培训。 当被问及他们在电子商务本地化中观察到哪些客户趋势时,许多受访者 (37%) 回答说需求正在增加。其他人确定了电子商务的具体趋势; 例如,一位受访者指出,印度的电子商务公司已经开始内部开展本地化活动。两名受访者观察到对更多语言的需求,而三名受访者则表示,电子商务本地化有使用更多机器翻译 (MTPE/PEMT或MT) 的趋势。 受访者被问及他们是否为任何OTT (over-the-top) 提供商工作,如海外流媒体服务、Netflix和亚马逊,或包括Hotstar和Viu在内的本地流媒体服务。 大多数受访者表示,他们目前不与这种类型的客户合作 (73%),而超过四分之一的人表示愿意 (27%)。 当被问及他们在电子学习中观察到什么客户趋势时,大多数 (65%) 通过他们的回答回应或暗示这一细分市场的需求正在增加。一些人限定了他们的回答。例如- 在电子学习中观察到的其他趋势包括更多的白话语言和一口大小的学习。 受访者被要求对5个特定需求驱动因素对其业务的重要性进行评分,1个 “根本不重要”,5个 “非常重要”。 受访者最重要的驱动因素是电子学习,超过92% 的人将其评为4或5。下一个最重要的驱动因素是电子商务,它获得了84% 的4s和5s。 规模的低端是国内金融科技和监管变化,被60% 的受访者评为4或5。 注: 调查参与者不一定回答所有问题。

以上中文文本为机器翻译,存在不同程度偏差和错误,请理解并参考英文原文阅读。

阅读原文