If you have already rewritten your CV for the jump between a startup and a large employer, you have done half the work. Two articles on this blog cover that half: one on reframing corporate experience for a startup CV, and one on adapting a startup CV when you move into a corporate role. Both of those are about a document that lists things. The cover letter is the document that argues, and the argument changes far more between a company of fifteen people and a company of twenty thousand than any CV bullet ever does.
Why the CV advice does not transfer
Editing a CV for a new audience is a sorting job. You decide what stays, what gets cut, which number goes first, whether "founding engineer" needs translating into a title a corporate reader recognises. It is mostly about accuracy of framing.
A cover letter has to do something a CV cannot. It addresses one specific person and gives that person a reason to move you forward this week. Who that person is, how much time they have, and what makes them nervous all change with the size of the employer. If your CV is already correctly reframed and your letter still reads identically in both directions, the letter is doing no work at all.
The person who opens it
At a startup, the letter usually lands with someone who would work beside you. A founder, a head of engineering, a first marketing hire now recruiting their own successor. They opened the hiring tool themselves, there might be sixty applications in the queue, and they are reading on a Sunday evening because hiring is one of nine things on their plate this month. Often no HR layer sits between you and them.
At a large employer there are at least two readers, and sometimes a machine before either of them. The first human is a recruiter or screener who will not manage you, who may be running eight open requisitions at once, and whose job is to produce a shortlist matching a brief someone else wrote. Your letter may be parsed by an applicant tracking system, attached to your profile, and left there until the hiring manager decides whether to open it.
A startup letter has one job: convince. A corporate letter has two: clear a filter written by someone who has never met you, then convince a manager who was not in the room for that first pass.
How long the letter actually survives
The startup reader is time-poor but genuinely reading. Three tight paragraphs, around 200 words, will be read end to end and sometimes forwarded verbatim to a co-founder. Padding is punished quickly because that reader recognises filler from their own inbox.
The corporate first pass is different. Twenty seconds, scanning for the job reference, the required qualification, notice period, location, and whether anything contradicts the CV. Your opening line has to name the role and the reference number, because the screener is holding several roles in their head and sorting is their actual task. The persuasion goes in paragraphs two and three, where the hiring manager will find it later.
Both letters fit on one page. Where the words go is not the same.
The evidence that lands on each side
A startup reader is trying to work out whether you can operate without scaffolding. What earns their attention:
- End-to-end ownership. You took something from nothing to live, including the unglamorous parts.
- Shipping speed with a date attached. "Live in six weeks" beats "successfully delivered".
- Breadth. You did the analytics and the support emails because there was no one else.
- Comfort with ambiguity. You made a call without a steering committee and it held.
A corporate reader is testing whether you can survive their machine. What earns their attention:
- Scale. Headcount, budget, number of markets, transaction volume.
- Stakeholder management. You moved something forward across functions that do not report to you.
- Process and compliance. Audit, regulatory approval, security review, procurement.
- Durability. You stayed and the thing you built is still running.
Both readers want results. They disagree about what a result is. One wants to know what you built and how fast. The other wants to know how you made it work across a matrix without breaking anything.
The same achievement, framed twice
Take one real fact. You replaced a billing system. Eight-month project, three people, and invoice errors dropped noticeably afterwards.
Written for a corporate screener and hiring manager: "I led the replacement of our billing platform across four European markets, coordinating finance, legal and two external vendors over eight months. The migration passed internal audit with no findings, and manual invoice corrections fell from roughly 400 per month to under 40."
Written for a founder: "Our billing was held together with spreadsheets and goodwill. I rebuilt it with two other people, shipped the first market in nine weeks, and cut manual invoice fixes by around 90 percent. I wrote the reconciliation scripts myself because we had no data team."
Same project. The first version proves you can run something across functions without creating a mess. The second proves you get your hands dirty and move. Notice what each one drops: the founder version never mentions the audit, and the corporate version never brags about writing scripts personally, because in a structured organisation that can read as a manager who cannot delegate.
Now run it the other way. Someone leaving a startup for a corporate role should not write "I did everything, we had no processes". Write instead: "I owned the full billing rebuild, wrote the runbook the finance team still uses, and handed over documentation before I left." That is startup breadth described in language a corporate reader trusts. Pulling your LinkedIn history into a structured CV with a tool like Postulit is useful here mostly because it gives you the raw dates and numbers you will reuse in both versions.
Tone, specificity, and what each reader treats as a red flag
With a startup you can be specific about the product, and you should be. Use it, notice something, say it plainly. An opinion is an asset: "your onboarding asks for a company size before it asks for an email, which is probably costing you signups" will get read twice. First person, plain sentences, no formulaic closing.
With a large employer, specificity moves up a level. Not the interface, but the division, the programme, the market position, the regulatory shift they are handling. There is already an article here on researching a company before writing a cover letter. The research is the same work; what you do with it differs. A corporate reader rarely welcomes unsolicited critique of their strategy, and a first-time applicant offering one reads as someone who will be difficult in a review cycle.
The red flags genuinely invert:
- Process language. A founder reading "I established governance frameworks and chaired the steering committee" hears overhead. A corporate screener reads the same line as competence.
- Casual tone. A founder finds it normal. A screener at a bank or an insurer reads it as someone who has not worked in a regulated environment.
- "I wore many hats." Fine at a startup if you name the hats. At a large employer it sounds like someone with no defined scope, which is hard to map to a job grade.
- Asking about structure. Questions about the career ladder or the review cycle are reasonable at a large employer and mildly alarming to a founder in month twelve.
- Stability signals. Eight consistent years at one employer is reassuring to one reader and, fairly or not, raises a question about appetite for mess in the other.
What to do this afternoon
Write the corporate version first, even if the startup role is the one you want. The corporate version forces you to be precise about scope, dates and numbers, and precision is hard to fake later. Then cut it down: strip the process scaffolding, drop about a third of the words, add one specific observation about the product, and replace the formal closing with a direct sentence about what you would work on in your first month.
Keep both versions in the same file. Most people apply in both directions across a career, and the second time you need one of these, the twenty minutes of rewriting is already done.