{"id":15465,"date":"2026-05-23T09:00:00","date_gmt":"2026-05-23T07:00:00","guid":{"rendered":"https:\/\/www.datamondial.com\/?p=15465"},"modified":"2026-05-06T15:48:17","modified_gmt":"2026-05-06T13:48:17","slug":"impact-inconsistent-data-annotation-ai-reliability","status":"publish","type":"post","link":"https:\/\/www.datamondial.com\/en\/impact-inconsistent-data-annotation-ai-reliability\/","title":{"rendered":"Why Inconsistent Data Annotation is Sabotaging Your AI Document Processing"},"content":{"rendered":"

The Hidden Costs of Human Variation in Data Training<\/h2>\n

When an Optical Character Recognition (OCR) model stalls during back-office operations, management often searches for technological causes. However, the underlying artificial intelligence rarely stops working on its own. Model failure is much more commonly the logical result of conflicting input from human operators during the training phase. Data validation for OCR, AI, and Machine Learning – DataMondial<\/a> shows that machine learning algorithms iteratively search for fixed, repeatable patterns within their assigned datasets. The moment this dataset contains internal contradictions, the algorithm becomes confused.<\/p>\n

On the shop floor, discrepancies quickly arise when bounding raw data. For example, when processing a scanned PDF, Operator A might highlight a gross weight including its unit of measurement (’25 kg’). Operator B, working the same shift, consistently records only the numerical value (’25’) on identical documents. To a human reader, this creates no difference in understanding. For a neural network, however, this variation immediately disrupts the extraction logic. The model cannot formulate a conclusive rule for what the specific ‘gross weight’ field actually entails. The direct result of this ambiguity is a spike in exception cases where the system demands human intervention.<\/p>\n

This problem is exclusively concentrated in unstructured data, such as scanned PDFs, commercial invoices, and physical waybills. With fixed Electronic Data Interchange (EDI) connections, where data is already highly structured via strict protocols, human variability in annotation does not occur. The real challenge lies in document streams where layouts fluctuate and contextual interpretation is required.<\/p>\n

Where the Interpretation of Logistics Documents Derails<\/h2>\n

Transport documents like customs declarations and waybills carry an inherent complexity. Layouts vary per freight forwarder, terminology is highly specialized, and data rarely sits at fixed coordinates. These variables inevitably trigger differences in human interpretation.<\/p>\n

A structural problem emerges from the variations in how composite company names are marked. One analyst might select ‘Maersk Logistics B.V.’, while a colleague extracts only ‘Maersk’, assuming the legal entity type is redundant for the operational process. The same inconsistency occurs when structuring addresses that are printed across multiple lines on the page. Should the postal code be merged into the street name field, or does it belong strictly with the city?<\/p>\n

The interpretation of Incoterms presents a similar hurdle. With the notation ‘FOB Rotterdam’, one data entry clerk might select the entire string as the delivery term. Another might label ‘FOB’ as the Incoterm and create a separate field for ‘Rotterdam’ as the location requirement. Without a strict frame of reference\u2014an established ‘ground truth’\u2014systems make random connections based on statistical chance. The algorithm lacks the guardrails to determine which operator followed the correct path.<\/p>\n

Practical Pitfalls at the Invoice Level<\/h3>\n

To remove the abstraction from this variability, the comparison below illustrates how two different analysts bound exactly the same line on a freight invoice differently within a labeling interface.<\/p>\n

\n

Line on the original scan:<\/strong>
\n04-11-2023 | Ocean Freight Shanghai – Spijkenisse incl. THC | \u20ac 1,450.-<\/em><\/p>\n<\/blockquote>\n\n\n\n\n\n\n\n\n\n\n
Data Field<\/th>\nAnalyst A Output (Detailed extraction)<\/th>\nAnalyst B Output (Grouped extraction)<\/th>\n<\/tr>\n<\/thead>\n
Date<\/strong><\/td>\n04-11-2023<\/td>\n04-11-2023<\/td>\n<\/tr>\n
Service Description<\/strong><\/td>\nOcean Freight<\/td>\nOcean Freight Shanghai – Spijkenisse incl. THC<\/td>\n<\/tr>\n
Origin<\/strong><\/td>\nShanghai<\/td>\nNo data selected<\/em><\/td>\n<\/tr>\n
Destination<\/strong><\/td>\nSpijkenisse<\/td>\nNo data selected<\/em><\/td>\n<\/tr>\n
Surcharges (THC)<\/strong><\/td>\nYes (boolean flag)<\/td>\nNo data selected<\/em><\/td>\n<\/tr>\n
Amount<\/strong><\/td>\n1,450<\/td>\n\u20ac 1,450.-<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Both outcomes are highly defensible from a human standpoint, but their conflicting structures prevent the AI from building a robust, predictive model for future ocean freight invoices.<\/p>\n

The Impact on Scalability in Back-Office Processes<\/h2>\n

The quality of source data correlates directly with the commercial outcomes of logistics operations. Inconsistent data training triggers a chain reaction that puts contract margins under severe pressure.<\/p>\n

The initial time savings gained from automated document extraction are instantly lost when outputs become unpredictable. Operations managers are forced to implement full manual checks (100% Quality Assurance) to prevent corrupt data from reaching the ERP or TMS. File turnaround times slow down, while operational expenditures (OPEX) rise to fund the headcount required for these secondary checks.<\/p>\n

This situation fuels a negative snowball effect within the ‘human-in-the-loop’ process. Employees who correct the AI’s mistakes during standard production feed these changes back into the system to make the model smarter. If these employees are working without strict annotation guidelines, they simply feed new deviations into the system. Existing model errors are thus sustained by conflicting back-end corrections. The result is a heavy retraining cycle that drains capacity away from processing your current live volumes.<\/p>\n

Moving Toward Uniform Annotation Guidelines<\/h2>\n

To eliminate the randomness of human input, a scalable data operation requires an architectural foundation built on strict annotation guidelines. Isolating individual thought processes forms the bedrock of this approach.<\/p>\n

This starts with comprehensively documenting edge cases. An operational manual shouldn’t just answer standard questions; it must provide a definitive ruling on irregular line breaks, merged table cells, and illegible stamps on freight documents. To safeguard the validity of the process, organizational segregation of duties is essential. The initial labeling of datasets must be completely decoupled from quality assessment. The person highlighting the data must never audit their own ‘ground truth’. To guarantee that the team subsequently operates as a single entity, data specialists quantify this uniformity using an objective metric.<\/p>\n

Measuring Inter-Annotator Agreement<\/h3>\n

Assessing uniformity is done via the Inter-Annotator Agreement<\/em> (IAA). This methodology, established within computational linguistics (as described by Artstein & Poesio (2008), “Inter-Coder Agreement for Computational Linguistics”, Computational Linguistics<\/em>), expresses the level of consensus among multiple reviewers as a concrete percentage or coefficient.<\/p>\n

The basic calculation simply looks at percentage overlap. If Rater A and Rater B independently assign labels to a sample of 100 invoice lines, and they draw exact bounding boxes around the exact same characters across 88 fields, the IAA score is 88%. In complex logistics extractions, the target is generally a minimum IAA of 95% before this trained data is allowed to flow into a neural network’s production environment. A drop in this figure immediately points to gaps in the underlying instructions or an individual gap in the operators’ domain knowledge.<\/p>\n

\n

Inconsistent data annotation disrupts the pattern-recognition capabilities of algorithms, driving up document processing turnaround times and inflating operational costs due to the need for continuous human correction. Establishing strict guidelines, combined with structured quality controls and measuring Inter-Annotator Agreement, forms the foundation for making document extraction truly scalable. Within complex logistics, e-commerce, and financial data streams, DataMondial serves as your specialized Dutch nearshoring partner based in Romania. By taking over Training AI models safely: The compliance checklist for data validation within the EU<\/a>, process knowledge, and a focus on Risk Reduction & Quality Assurance, DataMondial transforms your operational bottlenecks into a robust, measurable, and scalable BPO (Business Process Outsourcing) operation. Contact us for a targeted analysis of your data validation for OCR, AI, and Machine Learning – DataMondial<\/a> needs.<\/p>\n","protected":false},"excerpt":{"rendered":"

Discover how inconsistent data annotation AI disrupts machine learning, hikes up OPEX, and why strict guidelines and Inter-Annotator Agreement are the solution.<\/p>\n","protected":false},"author":10,"featured_media":15463,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_yoast_wpseo_focuskw":"","_yoast_wpseo_title":"The Impact of Inconsistent Data Annotation AI on Operations","_yoast_wpseo_metadesc":"When OCR and machine learning stall, the culprit is often inconsistent data annotation AI. Learn how to standardize data labeling for scalable BPO processes.","footnotes":""},"categories":[88],"tags":[],"class_list":["post-15465","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog"],"acf":[],"yoast_head":"\nThe Impact of Inconsistent Data Annotation AI on Operations<\/title>\n<meta name=\"description\" content=\"When OCR and machine learning stall, the culprit is often inconsistent data annotation AI. Learn how to standardize data labeling for scalable BPO processes.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"The Impact of Inconsistent Data Annotation AI on Operations\" \/>\n<meta property=\"og:description\" content=\"When OCR and machine learning stall, the culprit is often inconsistent data annotation AI. Learn how to standardize data labeling for scalable BPO processes.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/\" \/>\n<meta property=\"og:site_name\" content=\"DataMondial\" \/>\n<meta property=\"article:published_time\" content=\"2026-05-23T07:00:00+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"1376\" \/>\n\t<meta property=\"og:image:height\" content=\"768\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Ralph van Es\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Ralph van Es\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"6 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#article\",\"isPartOf\":{\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/\"},\"author\":{\"name\":\"Ralph van Es\",\"@id\":\"https:\/\/www.datamondial.com\/#\/schema\/person\/5438b776538ac7702fbaa3b85ebf463e\"},\"headline\":\"Why Inconsistent Data Annotation is Sabotaging Your AI Document Processing\",\"datePublished\":\"2026-05-23T07:00:00+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/\"},\"wordCount\":1133,\"publisher\":{\"@id\":\"https:\/\/www.datamondial.com\/#organization\"},\"image\":{\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg\",\"articleSection\":[\"Blog\"],\"inLanguage\":\"en-US\"},{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/\",\"url\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/\",\"name\":\"The Impact of Inconsistent Data Annotation AI on Operations\",\"isPartOf\":{\"@id\":\"https:\/\/www.datamondial.com\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#primaryimage\"},\"image\":{\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#primaryimage\"},\"thumbnailUrl\":\"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg\",\"datePublished\":\"2026-05-23T07:00:00+00:00\",\"description\":\"When OCR and machine learning stall, the culprit is often inconsistent data annotation AI. Learn how to standardize data labeling for scalable BPO processes.\",\"breadcrumb\":{\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#primaryimage\",\"url\":\"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg\",\"contentUrl\":\"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg\",\"width\":1376,\"height\":768,\"caption\":\"Two analysts labeling the same document differently due to inconsistent data annotation AI on a split screen.\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\/\/www.datamondial.com\/en\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Why Inconsistent Data Annotation is Sabotaging Your AI Document Processing\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.datamondial.com\/#website\",\"url\":\"https:\/\/www.datamondial.com\/\",\"name\":\"DataMondial\",\"description\":\"\",\"publisher\":{\"@id\":\"https:\/\/www.datamondial.com\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.datamondial.com\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Organization\",\"@id\":\"https:\/\/www.datamondial.com\/#organization\",\"name\":\"DataMondial\",\"url\":\"https:\/\/www.datamondial.com\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\/\/www.datamondial.com\/#\/schema\/logo\/image\/\",\"url\":\"https:\/\/www.datamondial.com\/wp-content\/uploads\/2022\/10\/datamondial_onderschrift.svg\",\"contentUrl\":\"https:\/\/www.datamondial.com\/wp-content\/uploads\/2022\/10\/datamondial_onderschrift.svg\",\"width\":431,\"height\":94,\"caption\":\"DataMondial\"},\"image\":{\"@id\":\"https:\/\/www.datamondial.com\/#\/schema\/logo\/image\/\"},\"sameAs\":[\"https:\/\/www.linkedin.com\/company\/datamondial\/\"]},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.datamondial.com\/#\/schema\/person\/5438b776538ac7702fbaa3b85ebf463e\",\"name\":\"Ralph van Es\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"The Impact of Inconsistent Data Annotation AI on Operations","description":"When OCR and machine learning stall, the culprit is often inconsistent data annotation AI. Learn how to standardize data labeling for scalable BPO processes.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/","og_locale":"en_US","og_type":"article","og_title":"The Impact of Inconsistent Data Annotation AI on Operations","og_description":"When OCR and machine learning stall, the culprit is often inconsistent data annotation AI. Learn how to standardize data labeling for scalable BPO processes.","og_url":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/","og_site_name":"DataMondial","article_published_time":"2026-05-23T07:00:00+00:00","og_image":[{"width":1376,"height":768,"url":"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg","type":"image\/jpeg"}],"author":"Ralph van Es","twitter_card":"summary_large_image","twitter_misc":{"Written by":"Ralph van Es","Est. reading time":"6 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#article","isPartOf":{"@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/"},"author":{"name":"Ralph van Es","@id":"https:\/\/www.datamondial.com\/#\/schema\/person\/5438b776538ac7702fbaa3b85ebf463e"},"headline":"Why Inconsistent Data Annotation is Sabotaging Your AI Document Processing","datePublished":"2026-05-23T07:00:00+00:00","mainEntityOfPage":{"@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/"},"wordCount":1133,"publisher":{"@id":"https:\/\/www.datamondial.com\/#organization"},"image":{"@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#primaryimage"},"thumbnailUrl":"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg","articleSection":["Blog"],"inLanguage":"en-US"},{"@type":"WebPage","@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/","url":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/","name":"The Impact of Inconsistent Data Annotation AI on Operations","isPartOf":{"@id":"https:\/\/www.datamondial.com\/#website"},"primaryImageOfPage":{"@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#primaryimage"},"image":{"@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#primaryimage"},"thumbnailUrl":"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg","datePublished":"2026-05-23T07:00:00+00:00","description":"When OCR and machine learning stall, the culprit is often inconsistent data annotation AI. Learn how to standardize data labeling for scalable BPO processes.","breadcrumb":{"@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#primaryimage","url":"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg","contentUrl":"https:\/\/www.datamondial.com\/wp-content\/uploads\/2026\/05\/impact-inconsistent-data-annotation-ai-reliability-en-featured.jpg","width":1376,"height":768,"caption":"Two analysts labeling the same document differently due to inconsistent data annotation AI on a split screen."},{"@type":"BreadcrumbList","@id":"https:\/\/www.datamondial.com\/de-impact-van-inconsistente-data-annotatie-op-de-betrouwbaarheid-van-ai-in-documentverwerking\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/www.datamondial.com\/en\/"},{"@type":"ListItem","position":2,"name":"Why Inconsistent Data Annotation is Sabotaging Your AI Document Processing"}]},{"@type":"WebSite","@id":"https:\/\/www.datamondial.com\/#website","url":"https:\/\/www.datamondial.com\/","name":"DataMondial","description":"","publisher":{"@id":"https:\/\/www.datamondial.com\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.datamondial.com\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Organization","@id":"https:\/\/www.datamondial.com\/#organization","name":"DataMondial","url":"https:\/\/www.datamondial.com\/","logo":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/www.datamondial.com\/#\/schema\/logo\/image\/","url":"https:\/\/www.datamondial.com\/wp-content\/uploads\/2022\/10\/datamondial_onderschrift.svg","contentUrl":"https:\/\/www.datamondial.com\/wp-content\/uploads\/2022\/10\/datamondial_onderschrift.svg","width":431,"height":94,"caption":"DataMondial"},"image":{"@id":"https:\/\/www.datamondial.com\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.linkedin.com\/company\/datamondial\/"]},{"@type":"Person","@id":"https:\/\/www.datamondial.com\/#\/schema\/person\/5438b776538ac7702fbaa3b85ebf463e","name":"Ralph van Es"}]}},"_links":{"self":[{"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/posts\/15465","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/users\/10"}],"replies":[{"embeddable":true,"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/comments?post=15465"}],"version-history":[{"count":2,"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/posts\/15465\/revisions"}],"predecessor-version":[{"id":15861,"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/posts\/15465\/revisions\/15861"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/media\/15463"}],"wp:attachment":[{"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/media?parent=15465"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/categories?post=15465"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.datamondial.com\/en\/wp-json\/wp\/v2\/tags?post=15465"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}