Yale SOM Consulting: One Word, Two Jobs, $60,000 Apart

The two consulting lines, and the gap between them

From the school's Class of 2024 employment report, checked 30 September 2026.

The median difference is $60,000. The 25th percentile difference is $62,500, and the 75th percentile difference is $42,000. So the two lines do not overlap much at all: internal consulting's 75th percentile of $150,000 sits well below external consulting's 25th percentile of $175,000.

One figure is identical across both lines, and it is the signing bonus at $30,000. Whatever separates the two jobs at Yale, it is base salary rather than the bonus convention.

The practical use is comparison hygiene. Most schools publish one consulting line. A candidate comparing Yale's number to another school's is comparing a split figure to a blended one, and the direction of the error depends on how much internal consulting the other school absorbed into its single row. Columbia is the only other program to split the word, which leaves external and internal consulting fused across the rest.

What Yale SOM says about its own consulting category

Yale warns about the word itself, in its own report: "More SOM students accepted their first role in consulting than in any other industry, but that category contains striking variety. Students accepted consulting roles that focus on the nonprofit industry, healthcare, and strategy among many other fields."

The same page reports that 5.2 percent of accepted jobs across all industries were in nonprofit or not-for-profit organizations, some of which sit inside the consulting row, and that 18 graduates started their own ventures.

That is a school telling readers its largest industry category is heterogeneous. The two salary lines above are the visible half of the heterogeneity. The invisible half is inside the external consulting line: a strategy firm and a nonprofit-focused consultancy both land there, and nothing in the table separates them.

So "consulting took more of the class than anything else at Yale" is true and thin. The useful question for a candidate is which of the several things inside that row they are actually recruiting for, and the answer has to come from the firm rather than the school.

The year the Yale SOM percentages disappeared into images

Yale published its shares as text through the Class of 2023 and stopped. The Class of 2023 data file prints them plainly.

For the Class of 2024, those shares exist only inside pie chart images on the web page. We downloaded all three charts and inspected them. The industry chart is 145,266 bytes and contains no text element whatsoever: every label has been converted to a vector path. Reconstructing a share from the geometry of an arc would be a fabrication, so we did not.

That has a consequence worth stating plainly. Any page quoting a Yale consulting percentage in 2026 is quoting the Class of 2023, whether or not it says so. This page says so.

The rounding in the 2023 file is worth a line too. The two function components are 43.5 and 10.8, which add to 54.3, while the document prints 54.2 for the total. The difference is rounding inside each line. We publish both readings rather than correcting the school's arithmetic on its behalf.

What Yale SOM says happened in 2024, without a number

Yale publishes a signed commentary on its own results. Its Assistant Dean writes: "Looking at the breakdown by industry, we see more students taking jobs outside the top five industries, primarily driven by large strategy consulting hiring normalizing after a few expansive years."

That sentence is the direction of travel, attributed, from the person who runs the careers office. It carries no figure, and the figures for that year are in the images. So the honest statement is that the school reports consulting hiring coming down from an unusual peak and does not quantify it in text.

The same commentary carries the most useful caveat any school in this batch prints. The school writes that this employment data "only reflects individual student choices" and says in the same sentence that it does not reflect their opportunities.

Read that against every share on this page. A 44.2 percent consulting figure describes what students accepted. It does not describe how many could have accepted, how many applied, or how many were rejected. Nobody publishes those, Yale included, and a page that presents a placement share as a measure of access has changed its meaning.

Offers, acceptances and the class that produced them

The Class of 2023 file gives the population behind the percentages, which most summaries skip.

The two rates are computed against the 294 seeking rather than the 336 in the class. So 269 of 294 is the 91.5 percent, and nine graduates who received an offer had not accepted one at the three month mark, because 269 minus 260 is nine.

Against the whole class, 260 acceptances out of 336 graduates is about 77 percent. That is a different and equally true number, and the difference between 88.4 and 77 is the 42 students who were not seeking employment. Every school publishes the first version. Only the count of the population lets you compute the second.

What the firms print about the case, and Yale does not

Yale publishes outcomes. The firms publish what happens in the room. Across seven firms we count 65 separate published instructions and 43 practice cases readable with no account, and a peer school's career library points its own students at nine firms' material on one page.

L.E.K. numbers seven steps for working a case and its first one is about the opening: "Listen to the question, repeat or paraphrase to be sure you understand; state your assumptions and restate the facts." Paraphrasing the question back is the cheapest error check available, and this is the firm that writes it down.

Accenture is the only firm in our corpus that publishes a duration, stating that "case interviews typically last from 45 to 60 minutes, with a focus on the case itself lasting 30 to 45 minutes". Both firms were checked 29 September 2026.

No firm publishes a weight. Across forty-four firm research files we hold, none contains a published scoring scheme, and that absence is the most reliable constant in the corpus. A page that shows you a firm's rubric with numbers on it is showing you something fabricated.

A plan for a school that publishes no dated calendar

Yale publishes no recruiting calendar with dates on the pages we could read. So the countdown method our Harvard recruiting page uses, where a 33 day gap between a deadline and a first interview sets the plan, is unavailable here. Diagnosis has to set it instead.

So run one short exercise of each type in a single evening, in the order below, and write down the dimension each one names as its first correction. Six readings in one sitting is the whole cost.

1. Math. One timed set, spoken aloud rather than worked silently.
2. Structuring. Ninety seconds on paper, then delivered without looking down.
3. Brainstorming. Breadth first, then whether the ideas fell into named buckets.
4. Exhibits. One chart read cold, with the caption covered.
5. Synthesis. One recommendation, answer first, with a risk attached.
6. Market sizing. One chain, with the unit stated before every product.

A full mock returns one impression of everything at once. Six short exercises return six named causes. Then work the weakest for a fortnight and re-run the whole diagnostic, which gives you a comparison rather than an impression.

For the length of a realistic practice block, the club recordings on our Yale case book page give a measured answer: four recorded mock cases averaging 29.1 minutes each.

Which dimension a Yale miss belongs to, and whose weights

The weights below are ours, read from src/lib/grading/rubrics.ts. No consulting firm publishes a scoring scheme, so weights attributed to a firm are invented.

Coverage carries 45 of the 100 points because an answer that misses a whole branch of the problem cannot be rescued by precision inside the branches it did cover. Concision carries 10 and is scored on whether the answer is short enough to deliver in one breath.

Run the structure drill and the math drill once each and read the flagged dimension. Every rubric and its reasoning sits on how an answer is assessed, and the drills are open in the tools with no account. What they do not do: no live interviewer, no business knowledge teaching, and none of our six assessment simulators is live today, which the assessments page states itself.

What Yale withholds, and who holds it

Five gaps, and the holder of each.

1. The Class of 2024 percentages in readable form. They exist as images on the school's page. The careers office holds the underlying table.
2. A hire count per firm. Yale publishes a list of hiring organizations by cycle and, in the pages we read, no counts against them.
3. A dated recruiting calendar. Nothing dated on the pages we read.
4. What its consulting club gives a member. We found no public club page on a Yale domain. We probed two student group platforms across six plausible identifiers and every address returned an empty application shell.
5. Whether the four club case recordings were authorized. They sit on a personal channel with 2016 upload dates, and neither the school nor the club claims them anywhere public.

We also decline to name alumni to approach. That belongs in the school's directory, and a list a stranger assembles from a blog post is the least effective version of it.

What changed on the Yale page, and what we deleted

30 September 2026. Rewritten after the school's employment report and its prior year data file were both read for the first time. The earlier version ran 372 words and told readers it would not report placement figures or name recruiting firms, because that belonged to the careers office and we had no way to verify it.

That was false in the way our process warns about most sharply. The school publishes salary percentiles in HTML, a full prior year in a PDF, and a signed commentary from its careers dean. Declining to look was presented as rigor. It was an absence of research, and the correction is the four tables above.

Two controls stay red on purpose. We hold no first-hand account from a Yale student describing their own consulting recruiting, because this session's search quota was spent before the page opened and the forum archives refuse our user agent. And we could not read the Class of 2024 shares, because the school vectorized them; rather than reconstruct a percentage from a pie slice we left the gap visible. Both are dated in the research file with the methods attempted.