{"course":{"slug":"ai-careers","title":"Working in AI","url":"https://addaly.com/learn/ai-careers","level":"beginner","track":"work"},"lesson":{"slug":"how-pay-actually-moves","title":"Raises, promotions and the market rate","url":"https://addaly.com/learn/ai-careers/how-pay-actually-moves","position":68,"of":85,"minutes":9,"answer":"Raises come from a small annual pool, promotions move your band, and job changes reset you to market, so if the band is the constraint, no amount of performance evidence will fix it internally.","keyIdea":"Raises come from a small annual pool, promotions move your band, and job changes reset you to market, so if the band is the constraint, no amount of performance evidence will fix it internally."},"body":{"format":"markdown","text":"## Three mechanisms, three different sizes\n\nPay moves in exactly three ways, and they are not interchangeable.\n\n**The annual raise.** A company sets a merit pool, commonly 3% to 8% of payroll in ordinary conditions, sometimes zero, and managers divide it across their team. Even an outstanding year usually means a share of a small pool. Your manager is not being unhelpful; they are dividing a fixed amount.\n\n**The promotion.** Moves you into a new band, and typically brings 8% to 15% at once. This is the only large internal step, and it is why the level lesson told you to negotiate the level rather than the number.\n\n**The job change.** Resets you to the current market. Typically 15% to 30%, and considerably more when you also change segment: services firm to product company, or one country's market to another's.\n\nThe sizes are the argument. Three years of good raises might be 18% compounded; one job change can be 25% in a day. This is not a recommendation to leave, and the next section explains why.\n\n> Figure: Three years of raises against one job change. Line chart of Pay, indexed to 100 against Years: Stay: a five per cent raise each year (0, 100) (1, 105) (2, 110) (3, 116) (4, 122) (5, 128) (6, 134); Change job in year three at plus 25 per cent (0, 100) (1, 105) (2, 110) (3, 138) (4, 145) (5, 152) (6, 160). A merit pool of three to eight per cent compounds to about eighteen per cent over three years; a change of job resets to market in a day. Not a recommendation to leave: the lesson goes on to what staying is worth.\n\n## Salary compression, which is a mechanism rather than an insult\n\nA company hires at market rate. Market rates rise. Existing staff receive raises from a pool that is smaller than the rise in market rate. Within three years, new joiners doing the same job earn more than people who have been there five years and are better at it.\n\nEvery large employer has this problem and very few fix it retrospectively, because fixing it costs a great deal all at once. Knowing the mechanism is worth more than resenting it: it tells you that loyalty is not rewarded automatically, that the correction is either a promotion or an external offer, and that your manager probably agrees with you and cannot do much.\n\n## Getting a raise, practically\n\n**Know the cycle.** Most companies decide pay one to three months before it is announced, and by then the pool is allocated. Ask your manager directly: *\"When are compensation decisions made, and when should I have made my case?\"* The number of people who ask is small and the answer is not secret.\n\n**Build the case over months, not in the conversation.** Keep a running file of what you delivered with numbers attached: the report that replaced four hours a week of manual work, the pipeline failure rate that fell from weekly to none, the analysis that changed a decision. Managers write these cases for their own managers and they are usually assembling them from memory. Handing over a page of evidence makes you the easy candidate.\n\n**Separate the two asks.** A raise and a promotion run on different machinery: one is a pool, one is a level change with a committee or a calibration meeting. Ask which you are eligible for and what specifically is missing, and get the answer as a behaviour rather than a feeling: *\"What would I need to be doing consistently for you to put me forward?\"*\n\n**Ask early and ask specifically.** \"I would like to discuss my compensation before the cycle closes; here is what I have delivered\" lands very differently from \"I feel underpaid\".\n\n## The counter-offer\n\nYour employer matches an outside offer to keep you. It is tempting and it is usually the wrong choice, for a reason more solid than the folklore: **the conditions that made you look are unchanged.** If you were leaving over the work, the manager, the ceiling or the commute, more money changes none of them, and you will be having the same thought in seven months with your bargaining chip spent.\n\nYou are also now known to have interviewed elsewhere, which affects how you are staffed and how you are considered when the next redundancy list is drawn up.\n\nYou will read confident statistics about most counter-offer acceptees leaving within a year. Those figures circulate widely and are poorly sourced; treat them as folklore. The structural argument above stands on its own and does not need them.\n\nThe exception worth naming: if you were happy and only underpaid, and the counter-offer fixes the band rather than adding a one-off bonus, staying can be entirely rational. Ask which of the two it is, in writing.\n\n## What staying is actually worth\n\nThe case for not changing job every eighteen months is real and rarely made:\n\n- **Scope compounds.** Senior work requires knowing a system and an organisation deeply, and you cannot demonstrate that in your first six months anywhere.\n- **Equity vests**, if you have any that is worth anything.\n- **Reputation accrues** among people who will hire you later: most senior roles are filled by people somebody already trusts.\n- **A CV of six jobs in seven years** raises a question you will be asked in every interview, and you will need a good answer each time.\n\nThe workable pattern for most people is two to four years per role early on, with a deliberate move when the learning stops rather than when the frustration peaks.\n\n## The limitation\n\nPay bands are set mostly by market, geography and employer type, and only marginally by how good you are. Excellent performance inside a low band produces a low-band raise, and no amount of evidence changes a band that was set by the company's business model.\n\nIf you are near the top of your band and the band is the problem, the fix is a different employer category, a different city, or a different market: the terrain module, revisited with better information than you had the first time."},"check":{"question":"An analyst is told they are a top performer but receives a 4% raise. They are near the top of their band. What follows from the mechanism?","options":[{"text":"Their manager undervalues them and the case should be made again with better evidence.","correct":false,"tempting":"Assumes advocacy failed, when a manager dividing a small pool may have given them the largest share available."},{"text":"A 4% raise is a normal share of a merit pool, and being near the top of the band means further internal movement requires a promotion into a new band, or a different employer, since bands are set by market and business model rather than performance.","correct":true},{"text":"They should obtain an external offer and use it to secure a match.","correct":false,"tempting":"May work once, but it spends their standing on a one-off correction that leaves the band (the actual constraint) unchanged."},{"text":"Top performers are usually paid through bonus rather than base, so the bonus is where the recognition will appear.","correct":false,"tempting":"Offers a plausible-sounding alternative mechanism that does not address being at the ceiling of a fixed range."}]},"prev":{"slug":"comparing-two-offers","title":"Two offers, compared with arithmetic","url":"https://addaly.com/learn/ai-careers/comparing-two-offers"},"next":{"slug":"freelance-economics","title":"Utilisation, and why a rate is not a salary divided by hours","url":"https://addaly.com/learn/ai-careers/freelance-economics"},"authorship":"Founder and series editor: Dr Mustafa Mun. Written by AI models to a written standard. Exam keys checked blind by a second model: 3,665 of 3,666 agreed. Academic review invited.","review":{"status":"Draft, open for academic review"},"cite":"Addaly (2026) 'Raises, promotions and the market rate', in Working in AI. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-careers/how-pay-actually-moves (Accessed: 13 October 2026).","licence":{"name":"CC BY-NC-SA 4.0","url":"https://creativecommons.org/licenses/by-nc-sa/4.0/"},"reader":{"schema":1,"title":"Raises, promotions and the market rate","courseTitle":"Working in AI","module":{"index":7,"slug":"the-money-and-the-paperwork","title":"The money, and the paperwork underneath it"},"answer":null,"keyIdea":"Raises come from a small annual pool, promotions move your band, and job changes reset you to market, so if the band is the constraint, no amount of performance evidence will fix it internally.","hero":{"kind":"figure","figure":{"t":"figure","label":"Figure","figure":{"type":"line","title":"Three years of raises against one job change","series":[{"label":"Stay: a five per cent raise each year","points":[[0,100],[1,105],[2,110],[3,116],[4,122],[5,128],[6,134]]},{"label":"Change job in year three at plus 25 per cent","points":[[0,100],[1,105],[2,110],[3,138],[4,145],[5,152],[6,160]]}],"xLabel":"Years","yLabel":"Pay, indexed to 100","caption":"A merit pool of three to eight per cent compounds to about eighteen per cent over three years; a change of job resets to market in a day. Not a recommendation to leave: the lesson goes on to what staying is worth."},"alt":"Figure: Three years of raises against one job change. Line chart of Pay, indexed to 100 against Years: Stay: a five per cent raise each year (0, 100) (1, 105) (2, 110) (3, 116) (4, 122) (5, 128) (6, 134); Change job in year three at plus 25 per cent (0, 100) (1, 105) (2, 110) (3, 138) (4, 145) (5, 152) (6, 160). A merit pool of three to eight per cent compounds to about eighteen per cent over three years; a change of job resets to market in a day. Not a recommendation to leave: the lesson goes on to what staying is worth."}},"blocks":[{"t":"h2","id":"three-mechanisms-three-different-sizes","runs":[{"text":"Three mechanisms, three different sizes"}]},{"t":"p","runs":[{"text":"Pay moves in exactly three ways, and they are not interchangeable."}]},{"t":"p","runs":[{"text":"The annual raise.","b":true},{"text":" A company sets a merit pool, commonly 3% to 8% of payroll in ordinary conditions, sometimes zero, and managers divide it across their team. Even an outstanding year usually means a share of a small pool. Your manager is not being unhelpful; they are dividing a fixed amount."}]},{"t":"p","runs":[{"text":"The promotion.","b":true},{"text":" Moves you into a new band, and typically brings 8% to 15% at once. This is the only large internal step, and it is why the level lesson told you to negotiate the level rather than the number."}]},{"t":"p","runs":[{"text":"The job change.","b":true},{"text":" Resets you to the current market. Typically 15% to 30%, and considerably more when you also change segment: services firm to product company, or one country’s market to another’s."}]},{"t":"p","runs":[{"text":"The sizes are the argument. Three years of good raises might be 18% compounded; one job change can be 25% in a day. This is not a recommendation to leave, and the next section explains why."}]},{"t":"h2","id":"salary-compression-which-is-a-mechanism-rather-than-an-insul","runs":[{"text":"Salary compression, which is a mechanism rather than an insult"}]},{"t":"p","runs":[{"text":"A company hires at market rate. Market rates rise. Existing staff receive raises from a pool that is smaller than the rise in market rate. Within three years, new joiners doing the same job earn more than people who have been there five years and are better at it."}]},{"t":"p","runs":[{"text":"Every large employer has this problem and very few fix it retrospectively, because fixing it costs a great deal all at once. Knowing the mechanism is worth more than resenting it: it tells you that loyalty is not rewarded automatically, that the correction is either a promotion or an external offer, and that your manager probably agrees with you and cannot do much."}]},{"t":"h2","id":"getting-a-raise-practically","runs":[{"text":"Getting a raise, practically"}]},{"t":"p","runs":[{"text":"Know the cycle.","b":true},{"text":" Most companies decide pay one to three months before it is announced, and by then the pool is allocated. Ask your manager directly: "},{"text":"”When are compensation decisions made, and when should I have made my case?”","i":true},{"text":" The number of people who ask is small and the answer is not secret."}]},{"t":"p","runs":[{"text":"Build the case over months, not in the conversation.","b":true},{"text":" Keep a running file of what you delivered with numbers attached: the report that replaced four hours a week of manual work, the pipeline failure rate that fell from weekly to none, the analysis that changed a decision. Managers write these cases for their own managers and they are usually assembling them from memory. Handing over a page of evidence makes you the easy candidate."}]},{"t":"p","runs":[{"text":"Separate the two asks.","b":true},{"text":" A raise and a promotion run on different machinery: one is a pool, one is a level change with a committee or a calibration meeting. Ask which you are eligible for and what specifically is missing, and get the answer as a behaviour rather than a feeling: "},{"text":"”What would I need to be doing consistently for you to put me forward?”","i":true}]},{"t":"p","runs":[{"text":"Ask early and ask specifically.","b":true},{"text":" “I would like to discuss my compensation before the cycle closes; here is what I have delivered” lands very differently from “I feel underpaid”."}]},{"t":"h2","id":"the-counter-offer","runs":[{"text":"The counter-offer"}]},{"t":"p","runs":[{"text":"Your employer matches an outside offer to keep you. It is tempting and it is usually the wrong choice, for a reason more solid than the folklore: "},{"text":"the conditions that made you look are unchanged.","b":true},{"text":" If you were leaving over the work, the manager, the ceiling or the commute, more money changes none of them, and you will be having the same thought in seven months with your bargaining chip spent."}]},{"t":"p","runs":[{"text":"You are also now known to have interviewed elsewhere, which affects how you are staffed and how you are considered when the next redundancy list is drawn up."}]},{"t":"p","runs":[{"text":"You will read confident statistics about most counter-offer acceptees leaving within a year. Those figures circulate widely and are poorly sourced; treat them as folklore. The structural argument above stands on its own and does not need them."}]},{"t":"p","runs":[{"text":"The exception worth naming: if you were happy and only underpaid, and the counter-offer fixes the band rather than adding a one-off bonus, staying can be entirely rational. Ask which of the two it is, in writing."}]},{"t":"h2","id":"what-staying-is-actually-worth","runs":[{"text":"What staying is actually worth"}]},{"t":"p","runs":[{"text":"The case for not changing job every eighteen months is real and rarely made:"}]},{"t":"ul","items":[[{"text":"Scope compounds.","b":true},{"text":" Senior work requires knowing a system and an organisation deeply, and you cannot demonstrate that in your first six months anywhere."}],[{"text":"Equity vests","b":true},{"text":", if you have any that is worth anything."}],[{"text":"Reputation accrues","b":true},{"text":" among people who will hire you later: most senior roles are filled by people somebody already trusts."}],[{"text":"A CV of six jobs in seven years","b":true},{"text":" raises a question you will be asked in every interview, and you will need a good answer each time."}]]},{"t":"p","runs":[{"text":"The workable pattern for most people is two to four years per role early on, with a deliberate move when the learning stops rather than when the frustration peaks."}]},{"t":"h2","id":"the-limitation","runs":[{"text":"The limitation"}],"role":"limit"},{"t":"p","runs":[{"text":"Pay bands are set mostly by market, geography and employer type, and only marginally by how good you are. Excellent performance inside a low band produces a low-band raise, and no amount of evidence changes a band that was set by the company’s business model."}]},{"t":"p","runs":[{"text":"If you are near the top of your band and the band is the problem, the fix is a different employer category, a different city, or a different market: the terrain module, revisited with better information than you had the first time."}]}],"check":{"question":[{"text":"An analyst is told they are a top performer but receives a 4% raise. They are near the top of their band. What follows from the mechanism?"}],"options":[{"text":[{"text":"Their manager undervalues them and the case should be made again with better evidence."}],"correct":false,"tempting":[{"text":"Assumes advocacy failed, when a manager dividing a small pool may have given them the largest share available."}]},{"text":[{"text":"A 4% raise is a normal share of a merit pool, and being near the top of the band means further internal movement requires a promotion into a new band, or a different employer, since bands are set by market and business model rather than performance."}],"correct":true,"tempting":null},{"text":[{"text":"They should obtain an external offer and use it to secure a match."}],"correct":false,"tempting":[{"text":"May work once, but it spends their standing on a one-off correction that leaves the band (the actual constraint) unchanged."}]},{"text":[{"text":"Top performers are usually paid through bonus rather than base, so the bonus is where the recognition will appear."}],"correct":false,"tempting":[{"text":"Offers a plausible-sounding alternative mechanism that does not address being at the ceiling of a fixed range."}]}],"order":[0,1,3,2],"why":null},"prev":{"slug":"comparing-two-offers","title":"Two offers, compared with arithmetic","minutes":10},"next":{"slug":"freelance-economics","title":"Utilisation, and why a rate is not a salary divided by hours","minutes":9}}}