Skai and the sale that would have happened anyway

Privalia had a perfectly reasonable plan. The fashion outlet already used search, social and app advertising. Display looked like the next place to grow. Its marketing team projected an initial €100,000 investment. There were models to support the idea. Then the company did something inconvenient: it tested whether the idea worked.

Using Skai’s Impact Navigator, Privalia set up a six-week experiment with geographically separate test and control groups. The new display channels produced no significant lift in the existing marketing mix. Privalia redirected the budget. A growth plan had failed its audition. That made the experiment useful.

The customer who was buying anyway

Advertising has a peculiar accounting problem. A person sees an ad, buys something, and the ad receives credit. But perhaps that person already wanted the product. Perhaps a promotion did the persuading. Counting the purchase is straightforward. Establishing what caused it requires a comparison with what would have happened without the advertising.

That distinction helps explain Skai, the company once called Kenshoo. It sells software for the work surrounding an advertising purchase: planning budgets, launching campaigns, adjusting bids, examining results and deciding what happens next. Retail media, paid search and paid social sit inside its remit.

Think of a brand advertising on Amazon, Google and Meta. Each system provides its own controls and reports. The brand still has to decide how the pieces work together. Skai supplies another layer, where teams can coordinate campaigns and put data to work across those boundaries.

Search was the beginning

Yoav Izhar-Prato, Nir Cohen and Alon Sheafer founded Kenshoo in 2006. The early business centered on search marketing. A $20 million growth round led by Bain Capital Ventures followed in 2014, with Sequoia Capital, Arts Alliance and Tenaya Capital participating. This was software built around an increasingly complicated buying process.

In 2021, Kenshoo acquired Signals Analytics and became Skai. The acquisition brought consumer and market intelligence into the mix. The name itself combines references to Signals, Kenshoo and actionable intelligence. Behind the wordplay was a broader ambition: connect the information informing a marketing decision with the machinery that carries it out.

Today, the homepage reports more than 8,200 brands. Public customer stories range from Bayer to Dr. Squatch; agencies are customers too. Skai competes in the same buying conversation as Pacvue, Perpetua and CommerceIQ. Its cross-channel approach matters most when a buyer needs to coordinate beyond a single retailer.

When the evidence moves the money

Bayer offers a more operational example. Working with Skai and measurement specialist Incremental, it sought additional growth for Aleve and One A Day. Incremental generated daily budget recommendations; an integration built by Skai Labs carried those recommendations into campaign workflows. Overall investment with the retail media network stayed flat.

Skai reports more than 1,000 daily budget optimizations and an improvement in incremental ROI exceeding 32%. The interesting feature is the connection between measurement and execution. A finding could change tomorrow’s allocation, repeatedly, without someone translating every recommendation into a manual edit.

Another experiment, with agency abaGada and Native Instruments, examined advertising around a winter promotion. Over four weeks, campaigns were switched off in a representative 30% of geographic markets. The analysis found advertising’s incremental impact was greater outside the days when promotional emails were sent. The teams changed their planning accordingly.

That is a lesson a reader can copy: specify the decision before choosing the metric. Are you deciding whether to enter a channel, increase a budget or change timing? Those are different questions. A reassuring total-sales number may answer none of them.

The six-figure coordination problem

Skai’s public pricing makes its intended customer unusually clear. Standard costs $114,000 annually for advertisers spending up to $4 million a year. Advanced is $276,000; Enterprise, $504,000; Enterprise Premier, $756,000. These are platform prices, separate from the media budget. Larger programs receive custom pricing.

The business model is enterprise SaaS, supported by training, customer success and services. The published tiers include different capabilities; incrementality testing appears in Enterprise Premier. Smaller programs are directed toward resellers. A sensible evaluation therefore asks which specific work disappears, which decisions improve, and whether that value covers the fee.

Its stated values include humility, innovation, passion and integrity. The employee gallery supplies a less abstract detail: colleagues playing music together. A workplace photograph cannot prove a culture, but it offers a welcome glimpse of people behind the optimization vocabulary.

The agent still needs a good question

Celeste AI arrived in April 2025, giving marketers a conversational route into analysis and recommendations. In May 2026, Skai announced Studio, where teams can build and coordinate specialized agents, with organizational guardrails and human approval checkpoints. Datahub supplies a separate offering for unifying marketing data.

“The future will be hybrid.”Gal Zohar, Skai chief data officer, September 2026

Agent Connect followed in September, opening Skai’s infrastructure to externally built agents. October brought a direct Uber Advertising integration across eight international markets and a Power Digital partnership connecting its Omega system to Agent Connect. Skai is selling access to the underlying data and actions as well as the interface.

There are practical limits. Geographic experiments need enough data and suitable comparison markets. A short test can answer a narrow question without settling the long-term value of a channel. And an agent needs reliable inputs, clear objectives and someone responsible for its decisions. Faster execution magnifies the value of asking the right question first.

Privalia’s experiment remains a useful way to understand the company. The attractive forecast did not survive contact with a test. The budget moved. For anyone buying advertising, that is a more useful outcome than another beautifully formatted report.