Yesterday Yelp closed at $18.96 and scored 79.77. Today it closed at $17.73 — down 6.5%, through its 52-week low — and scored 80.18. It is still first, and it moved further ahead of the field. That is not a fault in the screen. It is the screen doing exactly what it was built to do, in public, and it is worth understanding before you read the rest.
Cheapness carries 40% of the weight here and confirmation only 10%. A lower price improves every cheapness measure at once: on Yelp, enterprise value fell from 5.8 times operating profit to 5.4, and from 2.9 times free cash flow to 2.7. Meanwhile the confirmation score — the one that asks whether the market agrees — was already 9 out of 100 and had almost nothing left to give up. So the screen became more convinced as the market became less. Whether that is the screen finding something or the screen catching a falling knife is not a question two days can answer, and this note is not going to pretend otherwise.
How to read the four numbers
Every company gets four scores, and each is a percentile against other companies in its own sector, on the same day. Cheapness 98 means the company looks cheaper than 98% of its sector on enterprise value against operating profit, free cash flow and sales, plus what it returns to shareholders. It does not mean cheap against the market — and for one name below that distinction is the entire story.
The four combine into one composite: cheapness 40%, quality 30%, safety 20%, confirmation 10%. To be ranked at all a company must trade above $3, be worth $300 million, and turn over $2 million of stock a day. Banks, insurers and property trusts are excluded, because enterprise-value measures do not describe those balance sheets honestly. Tonight 1,181 companies were ranked and 154 were left out for want of data.
The five
| # | Company | Price | Score | Cheap | Qual | Safe | Conf |
|---|---|---|---|---|---|---|---|
| 1 | YELPYelp Inc · Consumer Services | $17.73 | 80.2 | 98 | 92 | 63 | 9 |
| 2 | NRDSNerdWallet · Technology | $8.10 | 79.5 | 98 | 84 | 65 | 23 |
| 3 | CRCTCricut · Industrials | $5.89 | 78.1 | 79 | 98 | 57 | 55 |
| 4 | CRLCharles River Labs · Biotech & Pharma | $277.33 | 77.4 | 86 | 74 | 67 | 75 |
| 5 | PGNYProgyny · Healthcare Services | $25.83 | 77.2 | 79 | 86 | 81 | 39 |
BorgWarner, fifth last night, has dropped out of the five. Progyny takes its place — it has been in the top ten since Monday and moved up as the others moved around it. Everything except the price comes from each company’s most recent SEC filings, which can lag a quarter behind the news, and for two of these five that gap matters.
1 · Yelp — cheaper by every measure, and still being sold
Cheapness of 98 out of 100, the highest reading on the board, on a business earning 24.7% on invested capital with net cash and negative accruals — profits backed by cash rather than by accounting. The share count is down 12.4% over the year.
Confirmation of 9. Down 24.9% over twelve months, trading at 70% of its 200-day average, and now below the 52-week low it was sitting on yesterday. Every point of cheapness it gained today it gained by falling.
Two things make the shareholder yield worse than it reads. It shows 29.4% tonight against 27.4% last night — but nothing new was returned to shareholders in a day. The figure rose because the market value it is divided by fell. And it is a trailing twelve-month number: Yelp paused its buyback on 6 August 2026, with $339m still authorised, to pay down a $100m revolving facility. It describes money already returned, at a price that no longer exists.
We found no company-specific news for Yelp on 22 or 23 September. It fell on two days when a number of consumer discovery and intermediary businesses also fell on fears about personal AI agents taking the middle out of their markets, though none of that coverage names Yelp. Which leaves the question the screen cannot answer: whether local advertising is cyclically weak or structurally exposed. Advertising revenue fell 3% last quarter, with restaurants, retail and other down 10%.
2 · NerdWallet — the same shape, the same warning
Cheapness of 97.6, effectively tied with Yelp for the best on the board, on a business earning 20.6% on capital with net cash and an eighth of its shares retired over the year. The accruals are negative here too.
Confirmation of 23, down 10.6% over the year at 78% of its 200-day average — and it fell again today. The cheapness sits on a revenue line under visible pressure: consumer credit-card revenue fell last quarter, which the company attributed to organic search traffic, while the small-business segment shrank 11%.
The top two names on this board are both businesses whose customers arrive through search, ranked in a week when the market was busy repricing exactly that risk. The screen cannot see that; it reads filings. Filings show two cheap, profitable, cash-generating companies. They do not show whether the channel that brings the customers still works in three years. That is the reading this list is sending you off to do.
3 · Cricut — the quality name, and the only one rising
Quality of 98.3, the highest reading anywhere in the five, on a 68.5% return on invested capital. It is the only one of the five above its 200-day average, at 1.26 times, and the only one of the eleven names this screen has followed since Monday that is up over the three days.
Safety of 57, the weakest in the group. And the quality is measured across a business whose halves move in opposite directions: last quarter products revenue — the machines and materials — fell 22% year on year while the subscription platform grew 5%.
A return on capital near 70% is real and not an accounting artefact. What a reader has to settle is what it is a return on: a contracting hardware base, plus a subscription attached to machines already sold. Both facts sit in the same filing.
4 · Charles River — cheap for biotech, not cheap
Confirmation of 75 on a stock up 94% over twelve months and trading 34% above its 200-day average. On the measure the other four struggle with, this is the strongest name here.
Thirty times operating profit is not cheap in absolute terms, and the 86 should not be read as if it were. It is cheap against other biotech and pharma companies, which is the only comparison this screener ever makes. If you want a name that is cheap against the market, it is not this one.
The trailing figures above describe a company that no longer exists in that shape. Charles River completed the sale of its CDMO and Cell Solutions businesses to GI Partners, and of several European discovery sites to IQVIA, both in May 2026 — together more than $200m off reported 2026 revenue. James Foster, who ran the company for decades, retired as chief executive on 5 May 2026; Birgit Girshick has the job. A screen reading the last twelve months of filings is partly measuring what was sold.
The open question is regulatory rather than financial. The FDA has stated an intention to move away from requiring animal testing in preclinical safety work, and research models and safety assessment are Charles River’s two largest segments. The company frames the shift as an evolution rather than a revolution. Whether it is, is the thing to go and read.
5 · Progyny — new to the five, and the steadiest of them
Safety of 81 and quality of 86, the best combination on the board. Net cash, a 24.8% return on invested capital, gross profitability of 0.45, and 8.9% of the shares retired over the year under a buyback the company renewed in May with $142.5m still to run.
Confirmation of 39 — the market is not arguing with it, but it is not agreeing either. And cheapness of 79 is the joint-lowest here: at nearly 20 times operating profit this is the least cheap name on a list sorted mostly by cheapness.
Progyny sells fertility and family-building benefits to self-insured employers, plus a pharmacy benefit. Reported revenue grew 5.3% last quarter — but 11% excluding one large employer that did not renew and whose transition revenue has now run out, which is why the headline understates the business underneath. Client count went from 542 to 604 over the year. What a reader should weigh is management’s own description of the market as “more brownfield than greenfield”: growth increasingly has to come from taking competitors’ accounts rather than from employers adopting the benefit for the first time.
What tonight actually showed
The most useful thing on this list is the thing that looks like an error. A company fell 6.5% and rose up a ranking, because the ranking weights cheapness four times as heavily as it weights the market’s opinion. If you think that is wrong, the honest response is not to quietly change the weights when a name goes against you — it is to write down in advance what would make you change them, and then test it. The weights here were fixed before the code that uses them was written.
Since Monday this screen has been recording the price of every company that reaches its top ten, from the day it arrives. There are three sessions of that now. It will be published when there is enough of it to mean something, which three days is not.
And the standing limit, which has not changed: in testing, this ranking ordered the universe — the bottom decile performed consistently worse than the top — but the top decile itself performed about in line with the market. A high rank buys a shorter reading list. It does not buy a return.
Sources: The Vance Report’s own screen file for 23 September 2026, built from SEC company facts and Nasdaq closing prices; the method is at how the Vance Value Screener works. Company events come from each company’s own second-quarter 2026 results and investor-relations releases: Yelp’s buyback pause; Charles River’s divestitures, leadership change and commentary on new approach methodologies; Cricut’s platform and product split; NerdWallet’s segment results; Progyny’s second-quarter results, client count and repurchase authorisation.