{"course":{"slug":"ai-careers","title":"Working in AI","url":"https://addaly.com/learn/ai-careers","level":"beginner","track":"work"},"lesson":{"slug":"the-numbers-of-a-search","title":"Running the search as a funnel","url":"https://addaly.com/learn/ai-careers/the-numbers-of-a-search","position":33,"of":85,"minutes":9,"answer":"A job search is a funnel with measurable conversion at each stage, and the route in changes the top-of-funnel rate by roughly ten times, which no amount of extra applications can compensate for.","keyIdea":"A job search is a funnel with measurable conversion at each stage, and the route in changes the top-of-funnel rate by roughly ten times, which no amount of extra applications can compensate for."},"body":{"format":"markdown","text":"## Feelings are a bad instrument\n\nThree weeks into a search everybody has the same two thoughts: *nothing is working* and *maybe I am not good enough*. Neither is information. A funnel is.\n\nEvery search has the same five stages, and the only useful question is which one is leaking.\n\n1. Applications sent or approaches made\n2. First conversations: recruiter or hiring manager\n3. Technical stages: take-home or live exercise\n4. Final stages\n5. Offers\n\n## The arithmetic, worked\n\nRates vary enormously by market, month and role, so treat these as a starting estimate to be replaced by your own within thirty applications. Typical shapes in an ordinary market:\n\n- **Cold applications → first conversation: 2% to 5%.** For heavily advertised remote roles it can be well under 1%, because a single posting can attract two thousand applicants.\n- **Referred or directly-approached → first conversation: 20% to 50%.** This is the entire argument of this module, expressed as a number.\n- **First conversation → technical: around 50%.**\n- **Technical → final: around 40%.**\n- **Final → offer: around 30%.**\n\nWork backwards from two offers, which is the right target because one offer is not a negotiating position:\n\n```\n2 offers  ->  ~7 finals  ->  ~17 technicals  ->  ~34 first conversations\n```\n\nThirty-four first conversations at a 3% cold rate is **1,130 applications**. At a 30% warm rate it is **113 approaches**.\n\nThose two numbers are the same search run two ways. One of them is possible for a person with a job and a life; the other is not. This is why the lessons on writing to strangers and on bridge jobs come before this one.\n\n> Figure: The same search, run two ways. Bars (people or applications needed at each stage): Cold applications at a 3% conversation rate 1130; Warm approaches at a 30% conversation rate 113; First conversations 34; Technical stages 17; Final stages 7; Offers 2. Working back from two offers at typical rates: 30% of finals convert, 40% of technicals reach a final, 50% of first conversations reach a technical. The top of the funnel is where the route decides everything, by a factor of ten.\n\n## Track it, in a spreadsheet, from application one\n\nSeven columns: date, company, role, route (cold, referral, recruiter, direct message), stage reached, outcome, next action and date. Nothing more. Ten seconds per row.\n\nAfter thirty rows you can answer the only question that matters (where does it break) and you will stop having to guess.\n\n## Reading your own funnel\n\n**Almost no first conversations.** The problem is upstream of your ability: targeting, route or the top third of your CV. If you are applying cold to roles two levels above your scope, no amount of technical improvement moves the number. Change route before changing yourself.\n\n**Conversations but no technical stages.** Something in the first fifteen minutes is not landing: usually a rambling answer to \"tell me about yourself\", or a salary expectation given badly. Both are fixable in an afternoon and neither is about your skills.\n\n**Technical stages that end there.** Now it is skills, and now, only now, the answer is more practice. Ask for the reason. About a third of companies will tell you something specific.\n\n**Finals without offers.** Rarely technical. Usually stories, evidence of judgement, or a stated concern nobody surfaced. Ask what the reservation was; at final stage people often answer honestly.\n\n## Volume is not the lever people think it is\n\nThe instinct when nothing happens is to send more applications. If your cold rate is 2%, doubling from 200 to 400 applications buys four extra conversations and about a hundred hours.\n\nThe same hundred hours spent making twenty warm approaches (a specific person, a specific reason, evidence attached) buys six. This is not motivational; it is division.\n\n## Two honest limitations\n\nFirst, these rates move with the market. In a tight hiring year the cold rate can fall below 1%, and a funnel that would have worked in 2021 produces nothing. That is not a statement about you, and confusing the two is what makes long searches damaging.\n\nSecond, a funnel measures throughput and says nothing about fit. Someone who accepts the wrong job efficiently has optimised the wrong thing. Keep a line in the tracker for why you wanted each role, and read it back after twenty entries. If most of them say \"it was open\", the search has drifted."},"check":{"question":"After 180 cold applications, Rahul has had four first conversations, three of which reached a technical stage and two of which reached a final. He has no offers. Where is his search actually leaking?","options":[{"text":"At the technical stage, since one of three conversations did not progress and that is where preparation pays off.","correct":false,"tempting":"Reads a 75% pass rate as the weak point, when it is well above typical and the sample is three."},{"text":"At the top: a 2% cold rate is normal, so the pipeline is starved rather than broken, and the fix is a warmer route (referrals and direct approaches) not better interviewing.","correct":true},{"text":"At the final stage, and he should focus on negotiation and closing.","correct":false,"tempting":"Draws a conclusion from two finals, where even a strong candidate converts around a third of the time."},{"text":"Nowhere; he should send another 180 applications, since the conversions are healthy.","correct":false,"tempting":"Correctly reads the conversions but chooses the lever with the worst return, buying four more conversations for another several months."}]},"prev":{"slug":"how-hiring-really-works","title":"How people actually get in","url":"https://addaly.com/learn/ai-careers/how-hiring-really-works"},"next":{"slug":"the-cv-that-survives","title":"A CV that survives a thirty-second read","url":"https://addaly.com/learn/ai-careers/the-cv-that-survives"},"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) 'Running the search as a funnel', in Working in AI. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/ai-careers/the-numbers-of-a-search (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":"Running the search as a funnel","courseTitle":"Working in AI","module":{"index":4,"slug":"getting-seen","title":"Getting seen: the routes that put a person in front of your work"},"answer":null,"keyIdea":"A job search is a funnel with measurable conversion at each stage, and the route in changes the top-of-funnel rate by roughly ten times, which no amount of extra applications can compensate for.","hero":{"kind":"figure","figure":{"t":"figure","label":"Figure","figure":{"type":"bars","unit":"people or applications needed at each stage","title":"The same search, run two ways","series":[{"label":"Cold applications at a 3% conversation rate","value":1130,"highlight":true},{"label":"Warm approaches at a 30% conversation rate","value":113,"highlight":true},{"label":"First conversations","value":34},{"label":"Technical stages","value":17},{"label":"Final stages","value":7},{"label":"Offers","value":2}],"caption":"Working back from two offers at typical rates: 30% of finals convert, 40% of technicals reach a final, 50% of first conversations reach a technical. The top of the funnel is where the route decides everything, by a factor of ten."},"alt":"Figure: The same search, run two ways. Bars (people or applications needed at each stage): Cold applications at a 3% conversation rate 1130; Warm approaches at a 30% conversation rate 113; First conversations 34; Technical stages 17; Final stages 7; Offers 2. Working back from two offers at typical rates: 30% of finals convert, 40% of technicals reach a final, 50% of first conversations reach a technical. The top of the funnel is where the route decides everything, by a factor of ten."}},"blocks":[{"t":"h2","id":"feelings-are-a-bad-instrument","runs":[{"text":"Feelings are a bad instrument"}]},{"t":"p","runs":[{"text":"Three weeks into a search everybody has the same two thoughts: "},{"text":"nothing is working","i":true},{"text":" and "},{"text":"maybe I am not good enough","i":true},{"text":". Neither is information. A funnel is."}]},{"t":"p","runs":[{"text":"Every search has the same five stages, and the only useful question is which one is leaking."}]},{"t":"ol","start":1,"items":[[{"text":"Applications sent or approaches made"}],[{"text":"First conversations: recruiter or hiring manager"}],[{"text":"Technical stages: take-home or live exercise"}],[{"text":"Final stages"}],[{"text":"Offers"}]]},{"t":"h2","id":"the-arithmetic-worked","runs":[{"text":"The arithmetic, worked"}]},{"t":"p","runs":[{"text":"Rates vary enormously by market, month and role, so treat these as a starting estimate to be replaced by your own within thirty applications. Typical shapes in an ordinary market:"}]},{"t":"ul","items":[[{"text":"Cold applications → first conversation: 2% to 5%.","b":true},{"text":" For heavily advertised remote roles it can be well under 1%, because a single posting can attract two thousand applicants."}],[{"text":"Referred or directly-approached → first conversation: 20% to 50%.","b":true},{"text":" This is the entire argument of this module, expressed as a number."}],[{"text":"First conversation → technical: around 50%.","b":true}],[{"text":"Technical → final: around 40%.","b":true}],[{"text":"Final → offer: around 30%.","b":true}]]},{"t":"p","runs":[{"text":"Work backwards from two offers, which is the right target because one offer is not a negotiating position:"}]},{"t":"code","lang":"","label":"Text","text":"2 offers  ->  ~7 finals  ->  ~17 technicals  ->  ~34 first conversations"},{"t":"p","runs":[{"text":"Thirty-four first conversations at a 3% cold rate is "},{"text":"1,130 applications","b":true},{"text":". At a 30% warm rate it is "},{"text":"113 approaches","b":true},{"text":"."}]},{"t":"p","runs":[{"text":"Those two numbers are the same search run two ways. One of them is possible for a person with a job and a life; the other is not. This is why the lessons on writing to strangers and on bridge jobs come before this one."}]},{"t":"h2","id":"track-it-in-a-spreadsheet-from-application-one","runs":[{"text":"Track it, in a spreadsheet, from application one"}]},{"t":"p","runs":[{"text":"Seven columns: date, company, role, route (cold, referral, recruiter, direct message), stage reached, outcome, next action and date. Nothing more. Ten seconds per row."}]},{"t":"p","runs":[{"text":"After thirty rows you can answer the only question that matters (where does it break) and you will stop having to guess."}]},{"t":"h2","id":"reading-your-own-funnel","runs":[{"text":"Reading your own funnel"}]},{"t":"p","runs":[{"text":"Almost no first conversations.","b":true},{"text":" The problem is upstream of your ability: targeting, route or the top third of your CV. If you are applying cold to roles two levels above your scope, no amount of technical improvement moves the number. Change route before changing yourself."}]},{"t":"p","runs":[{"text":"Conversations but no technical stages.","b":true},{"text":" Something in the first fifteen minutes is not landing: usually a rambling answer to “tell me about yourself”, or a salary expectation given badly. Both are fixable in an afternoon and neither is about your skills."}]},{"t":"p","runs":[{"text":"Technical stages that end there.","b":true},{"text":" Now it is skills, and now, only now, the answer is more practice. Ask for the reason. About a third of companies will tell you something specific."}]},{"t":"p","runs":[{"text":"Finals without offers.","b":true},{"text":" Rarely technical. Usually stories, evidence of judgement, or a stated concern nobody surfaced. Ask what the reservation was; at final stage people often answer honestly."}]},{"t":"h2","id":"volume-is-not-the-lever-people-think-it-is","runs":[{"text":"Volume is not the lever people think it is"}]},{"t":"p","runs":[{"text":"The instinct when nothing happens is to send more applications. If your cold rate is 2%, doubling from 200 to 400 applications buys four extra conversations and about a hundred hours."}]},{"t":"p","runs":[{"text":"The same hundred hours spent making twenty warm approaches (a specific person, a specific reason, evidence attached) buys six. This is not motivational; it is division."}]},{"t":"h2","id":"two-honest-limitations","runs":[{"text":"Two honest limitations"}]},{"t":"p","runs":[{"text":"First, these rates move with the market. In a tight hiring year the cold rate can fall below 1%, and a funnel that would have worked in 2021 produces nothing. That is not a statement about you, and confusing the two is what makes long searches damaging."}]},{"t":"p","runs":[{"text":"Second, a funnel measures throughput and says nothing about fit. Someone who accepts the wrong job efficiently has optimised the wrong thing. Keep a line in the tracker for why you wanted each role, and read it back after twenty entries. If most of them say “it was open”, the search has drifted."}]}],"check":{"question":[{"text":"After 180 cold applications, Rahul has had four first conversations, three of which reached a technical stage and two of which reached a final. He has no offers. Where is his search actually leaking?"}],"options":[{"text":[{"text":"At the technical stage, since one of three conversations did not progress and that is where preparation pays off."}],"correct":false,"tempting":[{"text":"Reads a 75% pass rate as the weak point, when it is well above typical and the sample is three."}]},{"text":[{"text":"At the top: a 2% cold rate is normal, so the pipeline is starved rather than broken, and the fix is a warmer route (referrals and direct approaches) not better interviewing."}],"correct":true,"tempting":null},{"text":[{"text":"At the final stage, and he should focus on negotiation and closing."}],"correct":false,"tempting":[{"text":"Draws a conclusion from two finals, where even a strong candidate converts around a third of the time."}]},{"text":[{"text":"Nowhere; he should send another 180 applications, since the conversions are healthy."}],"correct":false,"tempting":[{"text":"Correctly reads the conversions but chooses the lever with the worst return, buying four more conversations for another several months."}]}],"order":[3,2,1,0],"why":null},"prev":{"slug":"how-hiring-really-works","title":"How people actually get in","minutes":8},"next":{"slug":"the-cv-that-survives","title":"A CV that survives a thirty-second read","minutes":8}}}