What to Watch

What to Watch

The earlier chapters left a company that is easy to describe and hard to price — near-saturated, converting roughly half its revenue to cash, and guided to about 4% growth. At about $22 the shares price that rate in. Whether the guide is a cyclical trough or a structural ceiling will not be settled by argument. It will be settled by a handful of dated disclosures over the next four to eight quarters. This chapter sets out the facts both readings share and the checkpoints that separate them.

The guide the price is built on

Doximity's own fiscal-2027 outlook is the anchor. Management guided full-year revenue of $664 million to $676 million — about 4% growth at the midpoint — with an adjusted EBITDA margin of 49%, down from 55%, as AI-compute spending continues [1]. It framed HCP digital pharma-ad demand as soft with limited visibility, market growth "at or below 5%," and 65% of subscription revenue booked at guidance — in line with its three-year average [2]. Net revenue retention, the metric most directly tied to that growth rate, was 109% in fiscal 2026, down from 119% and 114% in the two prior years [3].

Sources: Q4 FY2026 earnings call, outlook [4]; FY2026 10-K, MD&A metrics [5].

The same facts, read two ways

The bull and the bear are not working from different numbers. They read the same disclosures in opposite directions. Setting the two readings side by side is the honest way to see where the disagreement actually lives — and it lives almost entirely in whether the deceleration is temporary.

Sources: Q4 FY2026 earnings call [6] and CEO Q&A [7]; FY2026 10-K, MD&A metrics [8] and SBC note [9]. The economics behind rows 3-5 are worked in The AI Contest and Margin of Safety.

Two facts on that table are worth separating from opinion because they are firmer than either narrative. NRR's mechanical link to the growth rate is the company's own description — it "fluctuates as that growth rate fluctuates" — so a retention rebound and a demand rebound are the same event, not two independent confirmations [10]. And the AI-Search TAM is a genuine adjacency, not the existing budget re-labelled: management places U.S. paid search near $19 billion and calls the healthcare-professional slice "incremental to the market we're in today" [11]. Both cut toward the bull, but neither is yet a number in the revenue line.

The checkpoints that separate the two

Each disagreement above resolves at a specific disclosure. The value of listing them is that most carry a date and a threshold, so the through-line's open question becomes checkable rather than rhetorical. Doximity's fiscal year ends March 31; the second half of fiscal 2027 is the October-2026 through March-2027 window, reported on the February-2027 and May-2027 calls.

Sources: Q4 FY2026 earnings call, outlook and CEO Q&A [12] [13]; FY2026 10-K, litigation note [14]; SEC comment letter, Feb 23 2026 [15].

The checkpoints do not arrive together, and the sequence matters. The SEC comment letter is the nearest: it was dated February 23, 2026, presses Doximity to define what "use" of its workflow tools means and to provide the metric for all periods, and runs on a ten-business-day response cycle — so how the engagement KPIs are revised is knowable soon, and those KPIs are the main non-financial evidence behind the AI-traction narrative [16]. Retention and booked-revenue conversion report every quarter. AI-Search revenue is a back-half fiscal-2027 event by management's own framing — it began selling in late April, expects no meaningful first-half contribution, and anticipates a more notable ramp into the fiscal back half [17]. The owner-FCF question does not clear until fiscal 2028, when stock-based compensation is guided to trend back down from the low-20s percent of revenue it reaches this year [18]. The OpenEvidence suit — now naming Doximity's Pathway Medical subsidiary and turning on Computer Fraud and Abuse Act data-access claims, with Doximity counterclaiming — has no scheduled resolution [19].

How the checkpoints map to value

The Margin of Safety chapter framed a fair-value range of roughly $15 in a structural-decline case, about $21 in a base case, and near $29 if a cyclical recovery and AI monetization arrive. The watch list is the bridge between those numbers and reality: the bear column filling out — NRR stalling, another sub-5% market year, AI Search staying minimal — pulls toward the low end, while the bull column confirming pulls toward the high end. What makes the setup asymmetric rather than merely cheap is underneath both scenarios: $749 million of net cash and a $493 million repurchase authorization that management has begun deploying near $22, against roughly $43 for the prior program [20]. That floor holds whichever way the demand question breaks; the checkpoints decide only how much sits on top of it.

None of the seven signals is decisive alone. The early, mechanical ones — the SEC metric definitions, quarterly retention, booked-revenue conversion — are the more informative: they arrive first and are harder to narrate around than the slower or noisier signals. The single reading that would most cleanly favor the structural case is a second consecutive year of sub-5% market framing with NRR still at or below 109%; the reading that would most cleanly favor the cyclical case is a disaggregated, material AI-Search line landing in the fiscal back half alongside a retention turn. Until those print, the deceleration is genuinely two-sided, and the disclosures above are where the evidence will come from.