In a digital-first world where milliseconds can make or break a sale, sluggish web performance remains a costly bottleneck for e-commerce brands. That’s the challenge Harper aims to solve with version 4.5, which delivers new capabilities for high-performance, data-intensive applications at scale.
First launched in 2017 as a distributed systems platform, Harper has evolved beyond its roots as a performance-focused database vendor into a full-stack application delivery platform — offering speed, scale, and simplified infrastructure to support high-performance digital operations.
The latest release introduces enhancements for building, scaling, and running high-performance workloads. Chief among them is support for binary large object (BLOB) storage, which enables the efficient handling of unstructured, media-rich content, such as images, video streams, and dynamic HTML.
According to Harper CEO and Co-founder Stephen Goldberg, the company started with the database layer to get the foundation right. Its evolution reflects a long-term vision to help digital commerce brands deliver faster user experiences while reducing IT costs and complexity.
Harper isn’t a CRM or an e-commerce platform. As Goldberg puts it, it’s “the plumbing” — the infrastructure that makes those systems faster and more cost-effective.
“In many of our deployments, we’re sitting directly in front of those e-commerce solutions or Oracle back / SAP back ends. As soon as the data is in Harper, the API, the caching layer, and the front end all have direct access in memory to that data as a single, unified platform,” he told the E-Commerce Times.
Unified Stack Optimizes Web Performance
Table of Contents
Harper built upon its earlier database technology to develop a unified approach to web performance. The result is a distributed clustering platform that deploys package schemas, endpoints, and application logic to a fleet of Harper instances, optimized for on-the-edge, scalable data delivery.
The platform accelerates the delivery of data-driven applications and websites by consolidating the traditional stack — API, caching, messaging, and database — into a single, unified system. The result: up to 7x faster page loads and 30x faster LCPs.
Network speed is critical for staying ahead of the competition — fractions of a second can make or break profitability. Truelist’s 2024 Shopping Cart Abandonment Statistics report notes that slow-loading websites can increase cart abandonment by 75%, and the global average cart abandonment rate stands at 75.6%.
Harper’s platform addresses the latency caused by backend complexity in traditional software stacks. This processing delay forces backend systems to stitch together databases, caches, application layers, and messaging queues, resulting in clogged network performance, increased serialization overhead, and operational complexity.
According to Jaxon Repp, CTO of Harper, BLOB storage is more than a new feature in the Harper platform. It is a core enrichment that opens up new use cases for the company’s customers.
“BLOB storage leverages Harper’s native streaming, sharding, and replication functionality to enable organizations to innovate without compromise, delivering unparalleled performance at any scale, for any type of data,” Repp explained.
Harper is seeing great results by caching full HTML pages as BLOBs, added the company’s SVP of Engineering, Kris Zyp. Some pages are a few hundred kilobytes.
“We’ve been able to update thousands per second, hitting network transfer limits. You can shard your data, replicate BLOBs in real time, and serve users from the closest geographic node for nearly limitless horizontal scale,” Zyp offered.
Architectural Scale Is Harper’s Key Edge
Other platforms have excellent products with great features, but Harper’s differentiator is its architectural design, Goldberg suggested.
Unlike other platforms, Harper is not in the market to build an end-user CRM or commerce solution. The company’s scope also touches other industries.
“Our technology is used in telecommunications. It’s used in defense to help identify enemy actors in the Pacific Ocean. It’s in one of the largest airlines in the world. Every flight update is sent from Harper,” Goldberg revealed.
Where commerce is concerned, Harper’s role focuses on significantly speeding up digital commerce use cases. Typical users of the platform include large enterprise development teams, and commerce use cases often involve web and SEO teams from some of the world’s largest retailers.
Goldberg noted that platforms like Squarespace can support vendors with anywhere from a handful of products to over 100, depending on scale.
“But if you’re talking about a retailer who has a billion SKUs with customers all over the planet, you’re reaching more of an enterprise-grade solution, and that’s where we provide a lot of that infrastructure,” Goldberg clarified.
“At our largest customer, we have 60 developers working on Harper. It’s typically companies where an e-commerce platform like Stripe or Vercel is scaled beyond that. So they built their own solution, and the dev team building that solution is our customer,” he explained.
AI Features and Commerce Use Cases Ahead
Goldberg is doubling down on enterprise retail, aiming to expand Harper’s use cases with solutions that elevate the digital commerce experience.
“I think where you’ll see us in the next couple of years is the entire digital commerce experience for many folks who roll their own. We’re already doing that with a few, but we plan to be the entire experience for most,” he envisioned.
Goldberg also plans to roll out some AI features, including vector search and semantic caching, in the near future. Harper already stores product data for its customers. New capabilities will let them search product data the way end customers do — using AI.
“That’s something many people do. Our unique value is that it’s slow. When things are slow, users leave your site,” he said to describe an ongoing retail problem that needs fixing.
Making Digital Infrastructure Faster, Cheaper
Today’s AI and large language models (LLMs) are not fast enough for the next level of solutions Harper will solve. Goldberg has no plans to rebrand again as an AI company.
“We’re not trying to be an AI company. We see ourselves as plumbers. So we’re going to make that plumbing significantly faster and cheaper so those can be more embedded in digital commerce experiences,” he said.
Goldberg noted that those new features are already under development. He alluded to releasing that plumbing fix sometime later this year.
