{"course":{"slug":"visual-storytelling","title":"Making It Say Something","url":"https://addaly.com/learn/visual-storytelling","level":"beginner","track":"create"},"lesson":{"slug":"what-success-looks-like","title":"Agree the measure before you make the thing","url":"https://addaly.com/learn/visual-storytelling/what-success-looks-like","position":8,"of":64,"minutes":8,"answer":"Agree one outcome measure and one diagnostic measure before production, build the attribution mechanism into the artwork itself, and state up front how noisy the result will be.","keyIdea":"Agree one outcome measure and one diagnostic measure before production, build the attribution mechanism into the artwork itself, and state up front how noisy the result will be, because a limitation named in advance is a position and the same sentence afterwards is an excuse."},"body":{"format":"markdown","text":"## The argument you will have in four weeks\n\nThe work goes out. Two weeks later the client says it did not work. You ask how they know. They say it feels quiet. You say the engagement was up. Neither of you is lying and neither of you can win, because nobody agreed in advance what the piece was supposed to do.\n\nThis argument is entirely preventable with one line written before you start.\n\n## Three kinds of measure, and which one is honest\n\n**Vanity measures**: views, likes, impressions, reach. They are easy to get and they move for reasons unconnected to your work: the algorithm, the hour of posting, one share by somebody with a large following. They are not useless; they tell you whether anybody saw the thing at all, which is the first of the three failure stages. But a piece with 50,000 views and no enquiries has failed, and a measure that calls it a success is the wrong measure.\n\n**Proxy measures**: saves, watch-through rate, time on page, scans of a QR code, clicks to the site. These sit closer to intent. Watch-through especially: for a short video, the proportion still watching at three seconds and at the end tells you where attention broke, which is diagnostic rather than merely score-keeping.\n\n**Outcome measures**: enquiries, walk-ins, bookings, sales, applications. This is what the client actually wants and the hardest to attribute, because a flyer, a reel, a friend's recommendation and the weather all landed in the same week.\n\nThe honest answer is to agree **one outcome measure and one diagnostic measure**, and to say plainly how confident attribution can be.\n\n## Make attribution possible while you are still designing\n\nThis is the bit that has to happen before production, not after, and it is almost always skipped.\n\n- A different phone number on the poster than on the website. Costs nothing on most systems, and it is the cleanest attribution available to a small business.\n- A specific phrase to mention: \"say KALYAN when you call\".\n- A separate short link or QR code per surface, so the flyer and the reel are distinguishable.\n- A one-line question from whoever answers the phone: \"where did you see us?\" Imperfect (people misremember), but it beats nothing and it is free.\n\nIf none of these are possible, say so now: \"we will not be able to tell this apart from everything else running that week, so we will judge it on enquiries during the period and accept that as rough\". A stated limitation before the fact is a professional position. The same sentence after the fact is an excuse.\n\n> Figure: Enquiries in six ordinary weeks, no campaign running. Bars (enquiries): Week 1 11; Week 2 15; Week 3 12; Week 4 16; Week 5 13; Week 6 14. Nothing was published in any of these weeks. The ordinary spread is about five, so a rise from twelve to fifteen after a flyer is inside the noise, and a campaign judged on it is being judged on the weather.\n\n## Write the number down, roughly\n\nNot because the forecast will be right (it will not) but because it turns a vague hope into something checkable. \"Last month, twelve enquiries. If this works, we would expect somewhere around twenty, not two hundred.\" Now everybody's expectations are in the same room before any money is spent.\n\nThis is also the moment to say if the target is unreachable by design work. If they need forty enquiries and the poster will be seen by roughly four hundred people, arithmetic says the campaign is too small regardless of how good the poster is. That conversation is much easier before you have made it.\n\n## The sample-size problem, in plain terms\n\nSmall businesses run small campaigns, and small numbers are noisy. The difference between eleven enquiries and fifteen is not evidence that a design worked; a normal quiet week and a normal busy week differ by about that much. You need either a larger volume or a longer period before a difference means anything, and the honest thing is to say which of the two you are relying on.\n\nBeware, too, of judging a change by the first day. A piece often does best in its first hours and then settles; comparing day one of the new work against a weekly average of the old is comparing two different things and will usually flatter the new.\n\n## What the measure cannot tell you\n\nA number tells you whether something happened. It does not tell you why, and it cannot distinguish a good piece badly distributed from a bad piece well distributed. Two additional habits close some of that gap:\n\n1. **Keep the diagnostic measure.** A three-second drop-off says the opening failed. A high watch-through with no clicks says the opening worked and the ask did not.\n2. **Collect three sentences of what people actually said.** \"I didn't realise you delivered\" is worth more than a percentage, because it names the belief that moved, and belief is what you were working on in the first place.\n\nWrite the measure into the brief, one line: *\"Success is enquiries mentioning KALYAN over four weeks; we expect around twenty, and we will read three-second drop-off to see whether the opening is doing its job.\"* That line ends the four-week argument before it starts."},"check":{"question":"A reel gets 80,000 views, a 4% watch-through to the end, and no enquiries. What does this pattern most usefully indicate?","options":[{"text":"The piece succeeded at reach and failed at conversion, so the call to action at the end needs to be stronger.","correct":false,"tempting":"Fixes the ending, but almost nobody reached the ending: a 4% completion rate means the failure happened long before the ask."},{"text":"The attribution was broken, so enquiries probably did arrive but were not counted.","correct":false,"tempting":"Possible but unsupported, and reaching for a measurement fault before reading the diagnostic figure is how teams avoid the finding."},{"text":"Attention was bought but not held, so the failure sits early in the piece.","correct":true},{"text":"The audience targeting was wrong, since the right audience would have watched to the end.","correct":false,"tempting":"Blames distribution for a retention curve that a mismatched opening produces just as readily, and it is the explanation that requires no change to the work."}]},"prev":{"slug":"using-reference-honestly","title":"Reference, and the line you actually cross","url":"https://addaly.com/learn/visual-storytelling/using-reference-honestly"},"next":{"slug":"one-message","title":"A design that says four things says none","url":"https://addaly.com/learn/visual-storytelling/one-message"},"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) 'Agree the measure before you make the thing', in Making It Say Something. Edited by M. Mun. Addaly. Available at: https://addaly.com/learn/visual-storytelling/what-success-looks-like (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":"Agree the measure before you make the thing","courseTitle":"Making It Say Something","module":{"index":1,"slug":"before-the-software","title":"Before you open the software"},"answer":null,"keyIdea":"Agree one outcome measure and one diagnostic measure before production, build the attribution mechanism into the artwork itself, and state up front how noisy the result will be, because a limitation named in advance is a position and the same sentence afterwards is an excuse.","hero":{"kind":"figure","figure":{"t":"figure","label":"Figure","figure":{"type":"bars","unit":"enquiries","title":"Enquiries in six ordinary weeks, no campaign running","series":[{"label":"Week 1","value":11},{"label":"Week 2","value":15},{"label":"Week 3","value":12},{"label":"Week 4","value":16},{"label":"Week 5","value":13},{"label":"Week 6","value":14}],"caption":"Nothing was published in any of these weeks. The ordinary spread is about five, so a rise from twelve to fifteen after a flyer is inside the noise, and a campaign judged on it is being judged on the weather."},"alt":"Figure: Enquiries in six ordinary weeks, no campaign running. Bars (enquiries): Week 1 11; Week 2 15; Week 3 12; Week 4 16; Week 5 13; Week 6 14. Nothing was published in any of these weeks. The ordinary spread is about five, so a rise from twelve to fifteen after a flyer is inside the noise, and a campaign judged on it is being judged on the weather."}},"blocks":[{"t":"h2","id":"the-argument-you-will-have-in-four-weeks","runs":[{"text":"The argument you will have in four weeks"}]},{"t":"p","runs":[{"text":"The work goes out. Two weeks later the client says it did not work. You ask how they know. They say it feels quiet. You say the engagement was up. Neither of you is lying and neither of you can win, because nobody agreed in advance what the piece was supposed to do."}]},{"t":"p","runs":[{"text":"This argument is entirely preventable with one line written before you start."}]},{"t":"h2","id":"three-kinds-of-measure-and-which-one-is-honest","runs":[{"text":"Three kinds of measure, and which one is honest"}]},{"t":"p","runs":[{"text":"Vanity measures","b":true},{"text":": views, likes, impressions, reach. They are easy to get and they move for reasons unconnected to your work: the algorithm, the hour of posting, one share by somebody with a large following. They are not useless; they tell you whether anybody saw the thing at all, which is the first of the three failure stages. But a piece with 50,000 views and no enquiries has failed, and a measure that calls it a success is the wrong measure."}]},{"t":"p","runs":[{"text":"Proxy measures","b":true},{"text":": saves, watch-through rate, time on page, scans of a QR code, clicks to the site. These sit closer to intent. Watch-through especially: for a short video, the proportion still watching at three seconds and at the end tells you where attention broke, which is diagnostic rather than merely score-keeping."}]},{"t":"p","runs":[{"text":"Outcome measures","b":true},{"text":": enquiries, walk-ins, bookings, sales, applications. This is what the client actually wants and the hardest to attribute, because a flyer, a reel, a friend’s recommendation and the weather all landed in the same week."}]},{"t":"p","runs":[{"text":"The honest answer is to agree "},{"text":"one outcome measure and one diagnostic measure","b":true},{"text":", and to say plainly how confident attribution can be."}]},{"t":"h2","id":"make-attribution-possible-while-you-are-still-designing","runs":[{"text":"Make attribution possible while you are still designing"}]},{"t":"p","runs":[{"text":"This is the bit that has to happen before production, not after, and it is almost always skipped."}]},{"t":"ul","items":[[{"text":"A different phone number on the poster than on the website. Costs nothing on most systems, and it is the cleanest attribution available to a small business."}],[{"text":"A specific phrase to mention: “say KALYAN when you call”."}],[{"text":"A separate short link or QR code per surface, so the flyer and the reel are distinguishable."}],[{"text":"A one-line question from whoever answers the phone: “where did you see us?” Imperfect (people misremember), but it beats nothing and it is free."}]]},{"t":"p","runs":[{"text":"If none of these are possible, say so now: “we will not be able to tell this apart from everything else running that week, so we will judge it on enquiries during the period and accept that as rough”. A stated limitation before the fact is a professional position. The same sentence after the fact is an excuse."}]},{"t":"h2","id":"write-the-number-down-roughly","runs":[{"text":"Write the number down, roughly"}]},{"t":"p","runs":[{"text":"Not because the forecast will be right (it will not) but because it turns a vague hope into something checkable. “Last month, twelve enquiries. If this works, we would expect somewhere around twenty, not two hundred.” Now everybody’s expectations are in the same room before any money is spent."}]},{"t":"p","runs":[{"text":"This is also the moment to say if the target is unreachable by design work. If they need forty enquiries and the poster will be seen by roughly four hundred people, arithmetic says the campaign is too small regardless of how good the poster is. That conversation is much easier before you have made it."}]},{"t":"h2","id":"the-sample-size-problem-in-plain-terms","runs":[{"text":"The sample-size problem, in plain terms"}]},{"t":"p","runs":[{"text":"Small businesses run small campaigns, and small numbers are noisy. The difference between eleven enquiries and fifteen is not evidence that a design worked; a normal quiet week and a normal busy week differ by about that much. You need either a larger volume or a longer period before a difference means anything, and the honest thing is to say which of the two you are relying on."}]},{"t":"p","runs":[{"text":"Beware, too, of judging a change by the first day. A piece often does best in its first hours and then settles; comparing day one of the new work against a weekly average of the old is comparing two different things and will usually flatter the new."}]},{"t":"h2","id":"what-the-measure-cannot-tell-you","runs":[{"text":"What the measure cannot tell you"}]},{"t":"p","runs":[{"text":"A number tells you whether something happened. It does not tell you why, and it cannot distinguish a good piece badly distributed from a bad piece well distributed. Two additional habits close some of that gap:"}]},{"t":"ol","start":1,"items":[[{"text":"Keep the diagnostic measure.","b":true},{"text":" A three-second drop-off says the opening failed. A high watch-through with no clicks says the opening worked and the ask did not."}],[{"text":"Collect three sentences of what people actually said.","b":true},{"text":" “I didn’t realise you delivered” is worth more than a percentage, because it names the belief that moved, and belief is what you were working on in the first place."}]]},{"t":"p","runs":[{"text":"Write the measure into the brief, one line: "},{"text":"”Success is enquiries mentioning KALYAN over four weeks; we expect around twenty, and we will read three-second drop-off to see whether the opening is doing its job.”","i":true},{"text":" That line ends the four-week argument before it starts."}]}],"check":{"question":[{"text":"A reel gets 80,000 views, a 4% watch-through to the end, and no enquiries. What does this pattern most usefully indicate?"}],"options":[{"text":[{"text":"The piece succeeded at reach and failed at conversion, so the call to action at the end needs to be stronger."}],"correct":false,"tempting":[{"text":"Fixes the ending, but almost nobody reached the ending: a 4% completion rate means the failure happened long before the ask."}]},{"text":[{"text":"The attribution was broken, so enquiries probably did arrive but were not counted."}],"correct":false,"tempting":[{"text":"Possible but unsupported, and reaching for a measurement fault before reading the diagnostic figure is how teams avoid the finding."}]},{"text":[{"text":"Attention was bought but not held, so the failure sits early in the piece."}],"correct":true,"tempting":null},{"text":[{"text":"The audience targeting was wrong, since the right audience would have watched to the end."}],"correct":false,"tempting":[{"text":"Blames distribution for a retention curve that a mismatched opening produces just as readily, and it is the explanation that requires no change to the work."}]}],"order":[0,3,2,1],"why":null},"prev":{"slug":"using-reference-honestly","title":"Reference, and the line you actually cross","minutes":9},"next":{"slug":"one-message","title":"A design that says four things says none","minutes":8}}}