4 ms·

The thing about advertising and marketing is that actually writing copy doesn't take anywhere near the full 40 hours a week. How much time does it take to type

The thing about advertising and marketing is that actually writing copy doesn't take anywhere near the full 40 hours a week. How much time does it take to type up a short bit of marketing copy? If your only goal is to just shit something, anything out then you could write up a whole campaign in a few minutes. But they don't do that in general. So where's ChatGPT going to save time? You'll still have to iterate, iterate, iterate, go back to stakeholders, discuss in progress work with clients, etc etc etc. I just don't see any reason to worry much in the short term, as long as you're working for a reputable company. This may change eventually, but I foresee it taking several years at least. Progress isn't anywhere near as "exponential" as some camps claim.

ineedasername 3y ago
Don't focus exclusively on the specific example I chose. It's only one area of applicability and tech is still advancing rapidly such that improvements will likely make it useful for new use cases as time goes on. But even staying on just this one example, your framing of the process is incorrect: You are underestimating the amount of "creative" time that goes on in these processes, and how the labor is divided. There are in fact full time creative folks who spend much closer to 40 hours a week on this than you think. I've worked on the analytics side of a few campaigns from small to very large (not my favorite type of project) so I've sat in on plenty of meetings where my organization was the customer. During the entire process, before we even sit down with a marketing firm, people on my side were spending significant time brainstorming ideas, slogans, narrative tone, visual aesthetic etc. so we could have a starting point for conversation. When that conversation began, it was account managers from the marketing firm, not creative staff involved. Account managers met with the client, discussed & clarified ideas, refined the parameters of what we wanted, etc. Then they took it back to their dedicated creative staff who did spend a majority of their time working on the actual creative side. We generally only spoke at length to someone like a full-time writer for very important projects, and their time was billed by the hour, usually scoped out as a block of time in the contract. Otherwise we might have only a brief conversation, usually if a few rounds of conversation & revisions hadn't quite got us where we needed. Sure there's bleed over and in any given company people will wear multiple hats and the venn diagram of account manage & creative will have more or less overlap, but beyond small firms this is a typical rough picture on the division of labor. >I just don't see any reason to worry much in the short term Depends on what you mean. If you mean "Don't panic that you're going to lose your job in a year" then I'd guess you're mostly right. If you mean "Don't worry at all" then you're mostly wrong, because as you specified, "short term". And as I said in my last comment, if people in jobs exposed to disruption aren't worried right now, then they are the ones most at risk of getting left behind. The people who worry now will be more likely to prepare. > your only goal is to just shit something, anything out... [and later] ...as long as you're working for a reputable company. See my previous comment about 90% of this output being fairly bland and mediocre, but the process still isn't to choose the first idea you think of, which is likely not going to be the best one compared to taking more time to come up with a bunch and choose from the best. There are lots of levels of mediocrity, and the first thing you shit out is simply the lowest of them. In reality the typical process will be something like this: Forget about longer copy writing, let's choose something smaller, a slogan. Usually 1 sentence, sometimes 2, but not much longer. I, the customer, go to a marketing firm for an entire campaign, one part of which is the slogan. As part of the contract we will have 2-3 rounds of discussions, preliminary ideas, revisions, etc, and the "deliverable" will be 3 finished, polished options to choose from. The account manager isn't going to their writer and saying "give me 3 slogans". They're saying "Give me 10 slogans to start with". The writer then goes to work. The writer does not write 10 slogans! The write writes more like 20 or 30, maybe more if you count minor iterations. They choose the 10 best and bring it to the account manager, talk them over, then the copywriter goes back and revises and maybe writes a bunch more for the account manager to finally choose the half dozen they think the client will like the most. The client gives feedback, the account manager goes back to the copywriter, who goes back for another round, etc. If all goes well then after 2 or 3 rounds of this the account manager has a final meeting with the client for approval of one of the 3 final options. And that's just for the slogan
- mrtranscendence 3y ago
  > And that's just for the slogan My point is that I don't understand what using an LLM changes here. You'll still have to generate multiple candidate slogans, you'll still have to bring 10 of them to the account manager, you'll still have to discuss the possibilities with the customer and tailor the results to their needs. You'll still have to iterate. All an LLM can do is write the candidate slogans, and -- as of now -- it doesn't do a spectacularly great job at it. Very little actual labor is being saved here.

ineedasername 3y ago

Don't focus exclusively on the specific example I chose. It's only one area of applicability and tech is still advancing rapidly such that improvements will likely make it useful for new use cases as time goes on. But even staying on just this one example, your framing of the process is incorrect: You are underestimating the amount of "creative" time that goes on in these processes, and how the labor is divided. There are in fact full time creative folks who spend much closer to 40 hours a week on this than you think. I've worked on the analytics side of a few campaigns from small to very large (not my favorite type of project) so I've sat in on plenty of meetings where my organization was the customer. During the entire process, before we even sit down with a marketing firm, people on my side were spending significant time brainstorming ideas, slogans, narrative tone, visual aesthetic etc. so we could have a starting point for conversation. When that conversation began, it was account managers from the marketing firm, not creative staff involved. Account managers met with the client, discussed & clarified ideas, refined the parameters of what we wanted, etc. Then they took it back to their dedicated creative staff who did spend a majority of their time working on the actual creative side. We generally only spoke at length to someone like a full-time writer for very important projects, and their time was billed by the hour, usually scoped out as a block of time in the contract. Otherwise we might have only a brief conversation, usually if a few rounds of conversation & revisions hadn't quite got us where we needed. Sure there's bleed over and in any given company people will wear multiple hats and the venn diagram of account manage & creative will have more or less overlap, but beyond small firms this is a typical rough picture on the division of labor. >I just don't see any reason to worry much in the short term Depends on what you mean. If you mean "Don't panic that you're going to lose your job in a year" then I'd guess you're mostly right. If you mean "Don't worry at all" then you're mostly wrong, because as you specified, "short term". And as I said in my last comment, if people in jobs exposed to disruption aren't worried right now, then they are the ones most at risk of getting left behind. The people who worry now will be more likely to prepare. > your only goal is to just shit something, anything out... [and later] ...as long as you're working for a reputable company. See my previous comment about 90% of this output being fairly bland and mediocre, but the process still isn't to choose the first idea you think of, which is likely not going to be the best one compared to taking more time to come up with a bunch and choose from the best. There are lots of levels of mediocrity, and the first thing you shit out is simply the lowest of them. In reality the typical process will be something like this: Forget about longer copy writing, let's choose something smaller, a slogan. Usually 1 sentence, sometimes 2, but not much longer. I, the customer, go to a marketing firm for an entire campaign, one part of which is the slogan. As part of the contract we will have 2-3 rounds of discussions, preliminary ideas, revisions, etc, and the "deliverable" will be 3 finished, polished options to choose from. The account manager isn't going to their writer and saying "give me 3 slogans". They're saying "Give me 10 slogans to start with". The writer then goes to work. The writer does not write 10 slogans! The write writes more like 20 or 30, maybe more if you count minor iterations. They choose the 10 best and bring it to the account manager, talk them over, then the copywriter goes back and revises and maybe writes a bunch more for the account manager to finally choose the half dozen they think the client will like the most. The client gives feedback, the account manager goes back to the copywriter, who goes back for another round, etc. If all goes well then after 2 or 3 rounds of this the account manager has a final meeting with the client for approval of one of the 3 final options. And that's just for the slogan

mrtranscendence 3y ago
> And that's just for the slogan My point is that I don't understand what using an LLM changes here. You'll still have to generate multiple candidate slogans, you'll still have to bring 10 of them to the account manager, you'll still have to discuss the possibilities with the customer and tailor the results to their needs. You'll still have to iterate. All an LLM can do is write the candidate slogans, and -- as of now -- it doesn't do a spectacularly great job at it. Very little actual labor is being saved here.