Summary

  • Yext's total annual recurring revenue increased from US$440.801 million at 30 April to US$440.815 million at 31 July 2026: just US$14,000, or 0.003%, at the company's disclosed precision.
  • Beneath that near-zero movement, ARR from customers at or above US$50,000 rose US$2.769 million while the below-US$50,000 book fell US$2.755 million.
  • Year over year, the larger book added US$6.556 million but the smaller book lost US$10.103 million, leaving total ARR down US$3.547 million.
  • Large-customer net retention improved to 98%, but it remains below the 100% line at which the starting cohort expands without help from new customers.
  • Corvo AI is a free, post-quarter experiment for small businesses. It cannot yet be counted as replacement ARR for the book Yext has deliberately allowed to contract.

A flat total produced by two moving parts

The first useful number in Yext's 1 September filing index is not revenue or profit. It is the US$14,000 difference between two quarter-end ARR balances. The Form 8-K furnished an earnings release, a shareholder letter and a separate product update for the quarter ended 31 July 2026. Together they show why an almost unchanged total can describe a business in active transition.

The shareholder letter reports total ARR of US$440.815 million, compared with US$440.801 million three months earlier. Customers carrying at least US$50,000 of ARR contributed US$405.880 million, up from US$403.111 million. Customers below that threshold contributed US$34.935 million, down from US$37.690 million.

The subtraction is unusually clean. The larger book gained US$2.769 million. The smaller book lost US$2.755 million. The two movements leave US$14,000. Total ARR therefore advanced by about 0.003%, even as the two components moved by roughly US$2.8 million in opposite directions.

That is not merely a rounding curiosity. It changes the question. A flat aggregate asks whether demand has stalled. The component bridge asks which customers Yext is building around and which customers it is prepared to lose.

The year-over-year view makes the direction harder to miss. ARR at or above US$50,000 increased from US$399.324 million to US$405.880 million, a gain of US$6.556 million, or 1.64%. ARR below US$50,000 fell from US$45.038 million to US$34.935 million, a loss of US$10.103 million, or 22.43%. Total ARR declined US$3.547 million. The larger book now supplies 92.07% of the total, against 89.86% a year earlier.

The preceding quarter's filing index and shareholder letter show that this was not a one-quarter reclassification. The below-US$50,000 balance descended through five disclosed quarter ends: US$45.038 million, US$43.212 million, US$40.622 million, US$37.690 million and US$34.935 million. The larger balance moved from US$399.324 million to US$405.880 million over the same endpoints, with one small sequential reversal between January and April.

The threshold is a measurement rule, not a customer biography

Yext calls the strategic destination enterprise. The table itself says something narrower: customers with ARR at or above US$50,000 and customers below US$50,000. It does not publish customer counts, legal size, employee count, location count or average contract value for either group. A customer under the threshold is not automatically a small business, and a customer above it is not automatically a multinational enterprise.

That distinction matters because retention uses a fixed historical designation. Yext starts with customers active twelve months earlier, assigns them to the cohort they occupied then and compares the same set's current ARR with its starting ARR. A customer that crosses US$50,000 during the year does not change cohorts inside that calculation. ARR acquired through a transaction also stays out of net and gross retention until one year after closing.

At July, dollar-based gross retention was 90% for the at-or-above-US$50,000 group and 69% for the smaller book. Net retention, which adds expansion back after churn and contraction, was 98% and 79% respectively. Total gross retention was 88%; total net retention was 96%.

The large book has improved. Its net retention rose from 96% a year earlier to 98%, and gross retention rose from 89% to 90%. But 98% remains below the self-compounding line. The customers present at the start of the measurement period still ended with 2% less ARR in aggregate after expansion, contraction and churn. New customers and other admissions outside that fixed cohort are therefore necessary to reconcile a sub-100% NRR with a 1.64% increase in the entire large-customer balance.

No exact new-logo dollar can be recovered by subtracting 98% from 101.64%. The two percentages use different populations. The retention denominator freezes a prior cohort and excludes recently acquired ARR; the balance comparison covers the whole book at each endpoint. The latest filed Form 10-Q before these results supplies operating and acquisition context, but it does not add the missing customer-level bridge for the July quarter.

Churn as a resource-allocation decision

Yext does not describe the smaller-book erosion as a surprise. It says it shifted its go-to-market strategy toward enterprise customers and is not dedicating material resources to mitigate churn in its current small-business base. Management says those exits reflect customers who were a poor fit for an enterprise product interface.

This is a control decision, not only a sales statistic. Product design, implementation labour, support queues and sales coverage can be optimized for complex accounts or for high-volume small accounts; doing both with the same interface and service model may destroy the economics of each. Yext has chosen to protect the enterprise path and let the old smaller book run down while testing a different delivery layer.

The strongest defence is visible in the numbers. Large-book growth accelerated from roughly 1% in the April quarter to 2% in July on the company's rounded basis. Gross and net retention both improved. The book reached 92% of total ARR. If those relationships carry higher lifetime value and lower service cost per dollar, losing mismatched small contracts can improve the quality of the aggregate even before total ARR returns to growth.

The limit is equally visible. A 90% gross-retention rate means the starting large cohort loses a meaningful share before expansion. A 98% NRR means expansion still does not restore all of it. Yext has improved the larger engine; it has not yet shown that this engine compounds on its own.

Corvo is an experiment, not a replacement ledger

Yext's answer to the smaller market is Corvo AI, a mobile-first conversational interface that can send business owners recommendations and execute marketing actions through natural language. The product announcement calls it an early version and makes it available free. It was released after the July quarter.

That timing creates a clean boundary. Corvo cannot explain the quarter-end ARR balance. It has no disclosed paid conversion, price, customer-acquisition cost, retention rate, support burden or ARR. It may eventually turn Yext's APIs, structured data and execution agents into a lower-cost small-business product, but today it is evidence of an option, not of recovered economics.

The same timing discipline applies elsewhere. Action Center reached general availability on 5 August, after quarter-end. Expanded brand-level Scout capabilities were still being piloted with a small number of enterprise customers ahead of planned broader availability on 30 September. GoShine was acquired in June and integrated as Brand Scout, but Yext says acquired GoShine revenue will not be material to consolidated results this fiscal year.

Those products may strengthen future retention and expansion. They do not belong in the US$440.815 million July balance unless the company says they were contracted and included at that date. Product availability, product use, signed ARR, recognised revenue and collected cash are five different receipts.

Profitability improved on a different clock

The earnings release reports quarterly revenue of US$111.103 million, down US$1.991 million or 1.76% year over year. Adjusted EBITDA rose from US$26.4 million to US$34.0 million, and its margin rose from 23% to 31%. Non-GAAP operating expenses fell to US$54.8 million from US$64.7 million.

That supports the efficiency case, but it does not turn ARR into profit. GAAP net income fell to US$13.1 million from US$26.8 million. The prior-year comparison contained a large favourable contingent-consideration adjustment that drove reported G&A below zero, so the GAAP expense swing is not a normal run-rate comparison. Operating cash flow of US$7.947 million and free cash flow of US$7.688 million were each slightly below the prior-year quarter.

The correct reading keeps the clocks separate. ARR measures annualized recurring contract value at an endpoint. Revenue measures services recognized during the quarter. Adjusted EBITDA removes specified costs. GAAP income keeps them but is distorted here by a prior-year acquisition adjustment. Cash flow records actual period movements. None can validate the others by substitution.

Sources