Understanding Hyperpersonalization

To use the analogy of a local pub, hyperpersonalization is not the same as the bartender getting regular customers their “usual.” Rather, it’s more like when the wait staff anticipates the customer’s next action based on their mood (jolly may mean celebration, gloomy signals just one drink); how they’re dressed (casual means here to stay, business suit says blowing off steam); what they have with them (groceries means quick visit, only a credit card means beer for everyone); or who they arrived with (friends means a big table, alone means at the bar). And all the while, the staff speaks the language that befits the circumstances.

The data that the wait staff sees and interprets to cater to your needs in the pub is exactly the kind that enterprises need to personalize their messaging to their customers. Companies and individuals — even in the same locale — exhibit differences in how they perceive, purchase, and use products and services, and these change depending on their journey and context.

Hyperpersonalization does not start and end with marketing; it is the way to engage with current and potential customers in product development, sales, billing, post-sales service, and more. For hyperpersonalization to work, content must be processed fast (just as the bartender sizes you up in seconds), enriched with abundant metadata (based on customer characteristics), and adjusted on the fly (to fit the purpose) across the various touch points of the end-to-end customer journey. The content must be connected to a wide array of systems and data sources, especially in a multilingual setting.

The Right Language Platform

The ability to assign the right translation modality based on content type and metadata is a defining feature of modern localization platforms. By aggregating and orchestrating different translation modalities — including the data needed to enrich them — across various content creation and delivery platforms, the modern TMS establishes itself as a central enterprise localization hub. With content connectors that enable application programming interface (API)-driven data exchange, they transcend the traditional “TMS” label to encompass upstream content tasks, becoming true language platforms.

In other words, this connectivity, aggregation, and orchestration are the strengths of modern localization platforms — or the saving grace for the TMS. Many in the industry forecast the downfall of traditional static and standalone TMS solutions. It’s hard to disagree with that, given the progress in content infrastructure and the increasing demand for automations and integrations in recent years. In a simplistic future scenario, plugging in an automated translation engine — say, a context-aware large language model (LLM) — and a simple (post-)editing interface into the content management system (CMS) solves the multilingual processing problem without needing a TMS.

However, such a solution only works if the CMS is the only client-side content repository — and even then, it doesn’t scale. Large enterprises with hyperpersonalization in mind need a localization platform that is scalable, is consistent across languages and locales, and has the ability to cater to the diverse multilingual content needs of the enterprise’s customers. This assumes connections to input systems beyond a CMS, including for pricing, product information management (PIM), digital asset management (DAM), enterprise resource planning (ERP), customer service, or e-commerce.

The Connectivity Challenge

The localization industry has come a long way when it comes to connecting all kinds of content repositories to TMSs. In the early 2000s, the industry made a first attempt to tackle the connectivity challenge by developing advanced file filters for different source file formats and aiming to standardize data exchange through the XML Localization Interchange File Format (XLIFF). Next, the focus shifted to automation and connectors, leading to a breakthrough in connecting translation business management systems (TBMSs) to computer-assisted translation (CAT) tools, a scenario typically benefiting the supply side of the localization spectrum. The latest developments are centered around enterprise language needs.

From an enterprise localization perspective, connectivity is often centered around CMSs, and the challenge seems relatively straightforward: Take any major CMS and automate data exchange with the TMS through APIs in a way that works for all CMS users. So why isn’t this the default approach for language work across all enterprise localization programs?

Some of the reasons are that many enterprises:

• build their own content repositories,
• heavily customize off-the-shelf CMSs, and/or
• have use cases that can’t be supported by traditional point-to-point connectors or require additional systems to be plugged in for the connection to produce the expected results.

Furthermore, many TMS providers develop APIs as an afterthought, often with insufficient documentation and priced without transparency, making it challenging for customers and users to make informed decisions. Automation is also often thought of as a productivity booster, rather than an enabler, and this decreases its priority and affects how its return on investment (ROI) is calculated.

Given the wide range of systems that take part in the wider localization process (spanning from request to billing) across a large enterprise, this approach is not scalable. Content sources such as CMSs, customer relationship management systems (CRMs), DAMs, and code repositories — as well as task management, business intelligence (BI), procurement, team collaboration, and user interface (UI) design tools — all take part in getting the right content delivered to current and potential customers. Linking each of these systems individually to a central language hub is a daunting task for localization programs if they need to be built from scratch. Due to the associated cost, time, and complexity, the envisioned holistic integration is oftentimes left at just that: a vision instead of reality, blocking the automations required for hyperpersonalization.

The Integrated Platform Approach

Certain vendors bring a fresh take on the problem with their API-first approach. This improves the integration experience by providing robust APIs accompanied by detailed documentation, providing the abstraction that technical staff unfamiliar with the translation management discipline may require.

Let’s take Phrase as an example. As one of the leading players in the TMS space — with a very high innovation capability, according to Nimdzi Insights’ “TMS Compass” analysis — Phrase offers a comprehensive and innovative automation and integration infrastructure with multiple layers. The foundational layer spans the entire platform, providing users with an extensive feature set to, for example, monitor content repositories, automate project creation, and manage continuous workflows for software keys or other specialized content types. The second layer includes a connector inventory with integrations for cloud storage buckets, version control systems, UI design tools, web CMSs, and marketing automation platforms. These integrations, comprising both “connectors” that pull content for translation and “plugins” that push content from third-party systems into the TMS, are designed to meet between 80 and 90 percent of typical integration needs. Finally, robust APIs and software development kits (SDKs) are available to deliver the final 10 to 20 percent of the typically required custom functionality.

This layered integration approach significantly reduces the complexity of creating integrated localization systems. Custom connector development time can be as little as 10 to 20 percent of what it would be without an integration layer, and connectors become more manageable and even portable. Even if the underlying APIs or systems are updated, the TMS vendor may perform this change transparently. With effort, time, and complexity barriers to connectivity lowered, enterprise localization teams can more confidently execute towards hyperpersonalized multilingual content flows.