Avensia fixes the parts of shopping you never see

A shopper sees a search box. A retailer sees thousands of opportunities to get the answer wrong. Products have variants, suppliers have their own descriptions, and the system that knows what is in stock may have a complicated relationship with the website. Avensia has built a business inside that gap: the distance between selling something online and making the machinery behave.

The Swedish company advises retailers, wholesalers and manufacturers, then helps implement the technology. Its work stretches from commerce strategy and user experience to product information, order management and integrations. The customer list includes Coop, Stadium, Dormy, Bygghemma and Piab. These are businesses for which a website is connected to warehouses, stores, purchasing agreements and years of accumulated operational habits.

That makes Avensia an interesting guide to the less photogenic side of digital retail. A redesign can be shown in a meeting. A cleaner data model is harder to applaud. Yet its customer work suggests that the second often determines how much the first can accomplish.

The search engine had to earn its traffic

At Coop Norway’s Obs and Obs BYGG, more than 30,000 online items made manual recommendations difficult to scale. The concern with upgrading search was specific: would Voyado’s newer engine understand an assortment containing building materials and DIY products well enough?

Avensia, Coop and Voyado ran the old and new engines alongside each other. During ten weeks of weekly reviews, the team assessed results and decided whether to send more traffic to Elevate 4. Initially, performance was similar. As the new engine learned, the reported results improved.

Those are separate metrics from separate sites, reported in Avensia’s customer case. Neither number represents a company-wide sales increase. The practical lesson is the rollout method: establish a baseline, watch real behavior and expand only when the evidence supports it. The experiment changed the customer’s confidence as well as its software.

Before the window dressing, 600,000 images

Stadium’s problem sat further upstream. Its existing product information management system had accumulated useful capabilities, but no longer fitted its technology plans. Moving to Bluestone PIM involved approximately 150,000 products and 600,000 images, supporting five sites. Seven integrations needed redesigning.

Avensia helped evaluate the platform, map requirements and coordinate the participants. It also developed a migration tool. Stadium supplied people who understood its data and workflows; Bluestone brought product expertise. The work was as much about preserving useful knowledge as replacing software.

Bygghemma faced a related tangle. Product and vendor information lived across a custom PIM and other systems, including its enterprise resource planning software. Avensia helped design a Syndigo master data management solution with clearer responsibilities and an integration layer. That structure was intended to make applications easier to change without rebuilding everything around them.

The distinction matters. PIM organizes information used to describe and sell products. MDM addresses shared business records more broadly. Either can become a bottleneck when nobody agrees which system owns which fact. A good-looking storefront cannot settle that argument.

A golf shop, a produce order, a family of factories

Dormy illustrates Avensia’s consumer-facing work. The golf retailer wanted an e-commerce platform that could support international expansion. Its stack combined Avensia Excite with commercetools, Contentful, Voyado and inriver. After weighing the expected benefits, Dormy brought the launch forward by three months, into golf’s peak season.

“We chose Avensia Excite because we needed a stable and scalable solution that could match our ambitious growth targets.”Anders Wall, CEO, Dormy

The relationship continued after launch. One subsequent feature translated staff expertise into a digital club selector. That is a useful picture of customer experience: helping someone make a choice that previously depended on a conversation in a shop.

Everfresh’s brief involved a different sort of urgency. The fresh-produce wholesaler serves around 1,300 B2B customers placing daily orders. Buyers may order goods before they arrive in stock. Substitutions, ordering cutoffs, contract prices and permissions all complicate the transaction. Avensia’s Nitro-based modernization addressed a platform that had become too inflexible for the business’s workflows and customer expectations.

At industrial automation company Piab, acquisitions had left brands with separate systems. Avensia used Nitro, Optimizely Cloud, inriver PIM and Azure Integration Services to build a shared foundation. Brands could move across gradually. Common templates and functionality gave the group a way to reuse development instead of commissioning the same work repeatedly.

Taken together, these projects explain the market Avensia occupies. Its customers need commerce to accommodate an existing business, with all its inconvenient details. A golf-club recommendation and a produce-order cutoff look unrelated on a screen. Both demand a developer who first understands the transaction.

Reusable parts, custom responsibility

Avensia’s Excite product packages preconnected components and shared code for composable commerce. The term describes an architecture in which functions such as content, commerce and search can be supplied by separate products. Excite uses Next.js and gives an implementation team a starting point rather than an empty repository.

Nitro is another part of the company’s offering, appearing in both retail and industrial customer projects. Around these products sits the larger service business: advice, architecture, development and continuing improvement. The value proposition is experience at the joints, where a platform’s standard behavior meets a customer’s unusual requirements.

This is competitive territory. Vaimo also offers commerce development, data management, optimization and accelerators. A joint commercetools and Contentful partner brochure presents Avensia Excite alongside Vaimo Launchpad and Valtech LEAP. Reusable building blocks are therefore a market category, not an Avensia monopoly.

Avensia’s particular case rests on the combination of commerce implementation and information management, backed by named projects. Buyers can assess that claim with concrete questions: Who will own the migration? Which integrations are already available? What changes after launch? Which members of the delivery team have handled an assortment or ordering model like ours?

Flexibility also carries obligations. This is an architectural tradeoff: choosing separate systems creates interfaces to maintain and decisions about responsibility. A business with a straightforward catalog and little internal technical capacity may gain less from that freedom. The relevant comparison is the total work required to operate the solution, including the customer’s own team.

Two birthdays and a decision to specialize

Avensia’s corporate history resists the tidy startup origin story. The listed group traces its roots to LUVIT, an e-learning venture begun at Lund University in 1998. The original Avensia software consultancy started in 2001. LUVIT acquired Avensia and INXL in 2007, creating a broader group.

The sharper change arrived in 2015. Selling Grade and Mashie left a commerce-focused business, and the group adopted the Avensia name. Chicago, Oslo and Gothenburg offices followed in 2018. In 2019, it acquired retail consultancy Insider Trends and expanded in the Philippines.

Gustafsson remains Group CEO. The management roster also includes Johan Liljeros as chief strategist and acting chief sales officer. His background spans business strategy, processes and technology, a combination that fits the company’s practical territory between commercial plans and implementation.

Avensia describes a workforce of around 300 and emphasizes global in-house teams, flexible remote work and work-life balance. These are the employer’s stated commitments. Its customer cases provide a more observable clue to its working style: joint teams, technical specialists and continuing relationships after the initial launch.

The economics behind the expertise

Avensia’s accounts reveal the cost structure behind those teams. Full-year 2025 net sales were SEK 423.3 million, almost unchanged from SEK 422.3 million in 2024. Operating profit rose to SEK 31.7 million from SEK 26.7 million. The company attributed improvement to a smaller cost base following workforce reductions; hourly rates also increased over the year.

The fourth quarter included SEK 3.1 million in charges for an Information Management savings program. This was Avensia’s restructuring cost, not a customer’s project bill. Treating the two as interchangeable would tell a very misleading story about implementation pricing.

In the first half of 2026, sales fell to SEK 206.9 million from SEK 228.0 million; operating profit declined to SEK 12.9 million from SEK 24.7 million. The second-quarter report described fewer billable hours and lower utilization. Personnel costs accounted for about 70 percent of first-half revenue. Expertise has a payroll even when customers postpone projects.

That is the useful economic lens for this company. A software accelerator can save development effort, but Avensia’s performance still depends heavily on deploying people productively. On August 19, the company also announced that CFO Lisa Logren had chosen to leave her position.

AI still has to pass the shop-floor test

Avensia’s April 2026 partnership with Constructor adds another route into AI-powered search and product discovery. A December 2025 agreement with Norce extends its B2B commerce partnerships. Both announcements broaden the company’s available tools; customer implementation and measured outcomes determine their eventual value.

The transferable lesson from Avensia’s work is procedural. Find a specific constraint, establish who owns the relevant data, keep a usable baseline and give the new system a test it can fail. That advice requires judgment: a lightly visited store may struggle to generate useful experimental evidence, while a complex migration requires customer participation that no accelerator supplies automatically.

For a buyer, the most revealing conversation may begin with one awkward detail: an unreliable product attribute, a contract price that will not display correctly, an order that arrives too late. Avensia’s case is strongest when it can follow that detail through the business and show how fixing it changes the sale.