News
Ulta Launched a Gemini Shopping Assistant and Its Ecom Comps Hit Mid-Teens: What Conversational Commerce Changes for Your DTC
Ulta posted Q1 FY26 net sales of $3.16B (+11.1%), comps up 5.3%. The split is the story: ecommerce comps grew mid-teens versus low-single-digit stores, and average ticket (+3.7%) drove far more growth than transactions (+1.6%). Its new Gemini-powered Ulta AI assistant is an AOV instrument, funded out of a 90 bps gross-margin gain while operating margin held flat near 14.2%.
Key Takeaways
- Ulta's Q1 FY26: net sales $3.16B, up 11.1% year over year, with comparable sales up 5.3%. Ecommerce comps grew mid-teens; physical-store comps grew low single digits. The digital channel grew roughly 4-6x faster than stores. SEC-verified against the 10-Q filed June 2, 2026.
- The growth came from bigger baskets, not more buyers. Average ticket rose 3.7% while transactions rose only 1.6%. Roughly 70% of the comp was basket size, not traffic. This is the post-2024 reality for almost every 8-figure DTC: traffic is flat and expensive, so the lever is AOV.
- Ulta AI, a Gemini-powered shopping assistant, launched in April 2026. It sits alongside a wider distributed-commerce build: a TikTok Shop storefront, Uber Eats across 1,500+ stores, and expanded Klarna. CEO Kecia Steelman called the early AI results promising. The pattern is conversational plus distributed, not one feature.
- The CFO move is hiding in the income statement. Gross margin expanded about 90 bps to 40.1%, but SG&A deleveraged about 80 bps to 25.7% of sales, holding operating margin roughly flat near 14.2%. Ulta funded the AI and channel build out of its margin gain, not out of operating income.
- For an 8-figure DTC, the takeaway is not build a Gemini agent. It is: split your comp into ecom-vs-other and ticket-vs-transactions, pick the one AOV lever you can actually move this year, and fund it from gross margin so your operating margin holds while you experiment.
On June 5, 2026, Digital Commerce 360 reported that Ulta Beauty credited a new AI shopping assistant for part of a strong digital quarter, and the underlying numbers matter for any 8-figure ecom operator because of where the growth came from. Ulta's Q1 FY26 net sales were $3.16 billion, up 11.1% year over year, with comparable sales up 5.3%. Ecommerce comps grew mid-teens while physical stores grew low single digits, and almost all of the comp was bigger baskets rather than more buyers. The company launched Ulta AI, built on Google Gemini Enterprise, in April. Below is what actually happened, why the basket-not-traffic split is the real signal, and what to watch next as AI-assisted discovery moves from headline to operating lever.
What happened
Ulta Beauty reported fiscal Q1 2026 results (the 13 weeks ended May 2, 2026) on June 2, and Digital Commerce 360 covered the AI angle on June 5, 2026. Net sales were $3.16 billion, up 11.1% from $2.85 billion a year earlier, a figure that matches the company's 10-Q on SEC EDGAR. Comparable sales rose 5.3%.
The channel split is the headline for operators. Ecommerce comparable sales grew in the mid-teens, while physical-store comparable sales grew low single digits. CFO Christopher DelOrefice confirmed the split on the call. CEO Kecia Steelman said fiscal 2026 was off to a strong start with broad-based growth, and that the sustained strength of the ecommerce channel is powered by the investments made over the last several years.
The growth was driven by basket, not traffic. Average ticket rose 3.7% and transactions rose 1.6%. The AI piece: Ulta launched Ulta AI, a shopping assistant built on Google Gemini Enterprise, in April 2026, live on Ulta.com and the mobile app. Steelman said initial results have been promising. It did not arrive alone. In the same window Ulta joined Uber Eats across 1,500-plus stores, launched a TikTok Shop storefront (a single April 16 livestream drew more than 5 million impressions), and expanded its Klarna buy-now-pay-later integration. Loyalty hit 47 million active members, up 4%. Ulta guided FY26 to net sales growth of 6-7% and comparable sales of 2.5-3.5%.
Why this matters for your business
Strip out the brand name and Ulta's quarter is a clean illustration of the two things that decide where 8-figure ecom growth comes from in 2026: the channel you fund, and the half of the comp equation you can actually move.
Start with the comp decomposition, because it is the most portable lesson. Comparable sales growth is always the product of two things: more transactions, or a bigger average ticket. Ulta's 5.3% comp broke down to transactions up 1.6% and average ticket up 3.7%. That means roughly 70% of the growth was basket size, not new buyers. This is not unique to beauty. For most 8-figure DTC brands right now, paid traffic is flat to declining in efficiency, organic is being reshaped by AI search, and the cheapest incremental revenue is the basket you already have in front of you. If you are not decomposing your own comp into traffic and ticket every month, you are flying blind on which lever is actually working. (Our average AOV by ecommerce vertical benchmark is a reasonable place to sanity-check where your ticket sits.)
That is the context for the AI assistant. An AI shopping assistant is, in operator terms, an AOV instrument. It earns its keep by helping a shopper find the right product faster (less abandonment), recommending real complements (more units per order), and answering specific requests your category navigation cannot. Ulta has 25,000-plus SKUs, where discovery is genuinely hard, so a conversational layer has obvious utility. The mid-teens ecommerce comp and the 3.7% ticket lift are consistent with that, even if no single quarter can prove the assistant caused it.
Then there is the move most people will skim past, which is the most useful one for a finance owner. Here is the year-over-year margin bridge from the filings:
| Metric (% of net sales) | Q1 FY25 | Q1 FY26 | Change |
|---|---|---|---|
| Gross margin | 39.1% | 40.1% | +0.9 pts |
| SG&A | 25.0% | 25.7% | +0.8 pts |
| Operating margin | ~14.1% | ~14.2% | ~flat |
Ulta expanded gross margin by about 90 basis points, then let SG&A rise by about 80 basis points, and operating margin came out roughly flat near 14.2%. That is a choice, not an accident. Ulta took the margin it gained on the product line and reinvested it into the channel, marketing, and technology build (the AI assistant, the new marketplaces, the app) instead of letting it fall to operating income. For a public company under pressure to show both growth and margin, holding operating margin flat while funding a heavy investment year is the disciplined version of "spend to grow." It is also exactly the framework an 8-figure DTC should use when it decides whether it can afford an AI or channel experiment: fund it from the margin you are improving, and treat operating margin as the line you do not cross.
The loyalty number reinforces the same point. Forty-seven million active members is the asset that makes the AOV lever work, because a logged-in, known member is who an AI assistant can actually personalize for. If your own loyalty or email list is underused, that is the cheaper version of the same play. (See loyalty member AOV lift by vertical for what that lift typically looks like.)
What to do this quarter
- Decompose your last four quarters of comparable sales into transactions and average ticket. If 60%-plus of your growth is coming from ticket, your roadmap belongs in AOV (bundling, recommendations, merchandising, AI discovery), not in more top-of-funnel spend. If it is coming from transactions, the opposite.
- Pressure-test whether AI-assisted discovery is a real lever for your catalog. It helps most when you have a wide assortment, a hard discovery problem, and a logged-in audience. If you sell 30 SKUs, an AI agent is mostly theater; spend on something else.
- If you test an AI search or recommendation layer, run it against a holdout, not against last year. The single biggest mistake here is crediting a tool for growth that loyalty, seasonality, or a merchandising change actually produced. A/B it.
- Build the experiment budget as a slice of your gross-margin improvement. Forecast the margin points you can realistically bank this year from sourcing, freight, returns, and mix, then size the AI and channel spend so operating margin holds flat. Use Ulta's bridge as the template: margin gained, margin reinvested, operating line protected.
- Audit your distributed-commerce surface. Ulta added TikTok Shop, Uber Eats, and expanded Klarna in one quarter. You will not do all of those, but pick the one marketplace or payment option where your customer already is and your competitors are not, and test it.
What we are watching next
Three signals over the next two quarters will tell us whether conversational commerce is a durable operator lever or a 2026 headline.
First, whether Ulta (and peers) start disclosing AI-attributed revenue or AOV lift with any specificity. Right now "initial results have been promising" is a sentiment, not a number. If Q2 or Q3 commentary puts a basis-point figure on AI-assisted basket lift, that becomes a real benchmark the rest of the market can plan against.
Second, the agentic-checkout question. As AI assistants from OpenAI, Google, and Perplexity move toward completing purchases inside the chat, the strategic risk for a retailer is that its own assistant matters less than whether its catalog is readable by someone else's agent. The brands that win the next phase are the ones whose product data, pricing, and availability are structured for machines to buy, not just humans to browse. That is the same muscle as being citable in AI search.
Third, the margin discipline holding. Ulta is guiding to slower second-half growth (6-7% net sales for the year against 11.1% in Q1). If the investment year continues and operating margin starts slipping rather than holding flat, that is the signal that the channel-and-AI build is getting expensive faster than it is paying back. Watch the SG&A line, not the press release.
The bottom line for 8-figure operators: the Ulta story is not "buy an AI chatbot." It is that growth in 2026 is a basket problem more than a traffic problem, that the tools to move the basket (AI discovery, loyalty, distributed channels) are real but have to be measured against a holdout, and that the only sustainable way to fund them is out of margin you are actually improving, with operating margin as the line you protect.
Sources and methodology
Primary news source. Ulta's AI assistant framing, the channel expansion details (TikTok Shop, Uber Eats, Klarna), the loyalty figure, and the executive quotes are from Digital Commerce 360, June 5, 2026. Comparable-sales splits (ecommerce mid-teens, stores low single digits), average ticket up 3.7%, and transactions up 1.6% are as reported in that coverage and on Ulta's earnings call.
SEC data. Net sales, gross profit, SG&A, operating income, and net income were pulled from Ulta Beauty's 10-Q filings on SEC EDGAR (CIK 0001403568) for the quarters ended May 2, 2026 and May 3, 2025. The margin percentages are computed from reported dollars: Q1 FY26 net sales $3,163,857K, gross profit $1,267,620K (40.1%), SG&A $814,699K (25.7%), operating income $448,256K (14.2%); Q1 FY25 net sales $2,848,367K, gross profit $1,114,219K (39.1%), SG&A $710,613K (25.0%), operating income $401,777K (14.1%). Basis-point changes are rounded.
Operator interpretation. The comp decomposition (roughly 70% of the 5.3% comp from ticket) is derived from the reported average-ticket and transaction growth rates. The framing of an AI assistant as an AOV instrument, and the "fund the experiment from gross margin" discipline, are Eightx operator interpretations applied to the public data, not statements by Ulta.
Limitations. Ulta does not break out ecommerce revenue in dollars or disclose AI-attributed revenue, so the ecommerce-vs-store comparison is on comparable-sales growth rates, and the AI contribution is directional. AOV dynamics vary widely by category; a beauty retailer's basket behavior will not map one-to-one onto apparel, supplements, or home goods.
Update cadence. We refresh this post when Ulta's Q2 FY26 results provide a second data point on the ecommerce-store gap or any AI-attributed disclosure, or when a major retailer puts a specific number on AI-assisted basket lift. Next scheduled review: end of Q2 FY26 earnings season (early September 2026).
Frequently asked questions
what is ulta ai and what does it actually do?
Ulta AI is a conversational shopping assistant Ulta launched in April 2026, built on Google's Gemini Enterprise platform and available on Ulta.com and the mobile app. In plain terms, it is a chat-style helper that lets a shopper describe what they want (a foundation for combination skin, a gift under $40, a dupe for a discontinued product) and get product recommendations they can buy without scrolling a category page. CEO Kecia Steelman said initial results have been promising. For operators, the important part is not the chatbot itself. It is that Ulta is using AI as a discovery and basket-building layer on top of a catalog of 25,000-plus SKUs, where the hard problem has always been helping a shopper find the right product fast.
why did ulta's ecommerce grow so much faster than its stores?
Ecommerce comps grew mid-teens while store comps grew low single digits, a gap of roughly 4-6x. Part of that is structural (ecommerce is a smaller base, so it compounds faster), and part is that Ulta has been pouring investment into digital for several years: site search, the app, the loyalty integration, and now the AI assistant and new channels like TikTok Shop and Uber Eats. The CEO was explicit that the ecommerce strength is powered by multi-year investment, not a one-quarter spike. The operator lesson is that the channel you fund is the channel that grows, and right now Ulta is funding digital.
does an ai shopping assistant actually raise aov?
It can, but the mechanism matters. An AI assistant lifts average order value when it does three things well: surfaces a relevant product faster than browsing (which reduces abandonment), recommends genuine complements rather than random cross-sells (which adds units), and handles the long tail of specific requests your category navigation cannot. Ulta's quarter is consistent with this: average ticket rose 3.7% against transactions up only 1.6%, so the basket did the work. That said, you cannot attribute all of that to the assistant, because Ulta also expanded loyalty, channels, and merchandising in the same quarter. Treat AI-assisted discovery as one AOV lever among several, and measure it against a holdout, not against last year.
should my 8-figure dtc build its own ai shopping agent?
Probably not a custom one, not yet. Ulta is a multi-billion-dollar retailer with the engineering budget to deploy Gemini Enterprise across 25,000 SKUs. For an 8-figure DTC, the higher-ROI version is usually an off-the-shelf AI search or recommendation app on your existing platform, pointed at your real conversion problem. Before you spend anything, answer one question: is your growth constrained by traffic or by basket size? If transactions are flat and AOV is your lever (as it is for most brands right now), an AI discovery or bundling layer is worth testing. If traffic is your constraint, the money belongs in acquisition or retention, not a chatbot.
how do i fund an ai or channel experiment without tanking operating margin?
Do what Ulta did. In Q1, Ulta expanded gross margin about 90 basis points but let SG&A rise about 80 basis points, so operating margin stayed roughly flat near 14.2%. In other words, it spent the margin gain on the build instead of promising the board both better margin and more investment. For an 8-figure brand, the discipline is the same: size the experiment as a share of the gross-margin improvement you can actually bank this year (from better sourcing, freight, returns, or mix), and hold operating margin as the constraint. That way a failed test costs you a margin point you were going to reinvest anyway, not a hole in the bottom line.