Almost Timely News: What ChatGPT is Really Good At, Measurement Strategies for Agencies Course

Warning: this content is older than 365 days. It may be out of date and no longer relevant.

Almost Timely News: What ChatGPT is Really Good At, Measurement Strategies for Agencies Course (2023-01-22) :: View in Browser

Almost Timely News

Get This: New FREE Course

I’m mildly excited to announce that we’ve got a new mini-course, and this one’s free. It’s called Measurement Strategies for Agencies. You’ll learn the 5 things agencies do most wrong when it comes to developing effective measurement strategies for clients – and how to fix it. It’s just about an hour long, it’s free, and it’s for two groups of people:˝

  1. People who work at agencies, so you get better at measurement
  2. People who HIRE agencies, so you know what to ask for in your reports

👉 Click here to take this course now for free!

Watch This Newsletter On YouTube 📺

Almost Timely News: What ChatGPT is Really Good At, Measurement Strategies for Agencies (2023-01-22)

Click here for the video 📺 version of this newsletter on YouTube »

Click here for an MP3 audio 🎧 only version »

What’s On My Mind: AI Creation Vs. Refinement

Let’s take a minute to talk about originality, AI, and content marketing. A lot of folks, myself included, have had a lot to say about generative AI, about how AI is ushering in a new age of generated content. Machines that write almost as well as we do on average, machines that can crank out incredible artwork.

Yet, that’s not what these machines excel at.

When it comes to the use of artificial intelligence, in particular, large language models like GPT-3 and the ChatGPT interface, what these models are good at is transforming inputs. Large language models, in general, are nothing more than massive statistical probability matrices. There was a great quote from the “This Week in Machine Learning and AI” podcast that goes something like “a word is told by the company it keeps.”

This means that these tools and models understand language only to the extent of the statistical distributions of the words, phrases, sentences, and paragraphs that they appear in. That’s why they can replicate grammar very well because grammar is nothing more than statistical distributions of words. They’re autocomplete on steroids.

For example, what’s the next word in these sentences?

“For all you do, this Bud’s for _____”

“That’s not a knife, this is a _____”

“God save the _____”

We know what these various sentences are because in their respective cultures, they’re so frequently used that we are accustomed to the word distributions, though the last one is now changing once again because the people it’s about have changed.

What this means is that from a generation capacity, these tools can generate text very capably, but that generation is going to be mathematically an average of the text that most commonly surrounds those keywords. That’s why your prompts to generate new stuff have to be so incredibly detailed, so that the tools can understand the increased sense of probabilities for the words you’re asking them to generate.

Telling a large language model to write a blog post about social media marketing is going to generate extremely bland, average content. Telling it to generate social media content about the engagement rates on TikTok with regard to time of day and gender is going to give you more specific content because the large language model itself can understand based on the additional words you’ve provided, more of the context. It is drawing from additional statistical probabilities from those words – a word is known by the company it keeps.

However, what these tools produce is still a statistical average of what they’ve been trained on. They’re not going to produce anything original because they can’t by definition. Certainly, they’ll produce original orderings of words to some degree, but they can’t produce new concepts that aren’t in the original model. That’s why it’s such a big deal when new versions of models – bigger models especially – get released, because the models have more original ideas in them to work with.

So some marketers are going to create an avalanche of average, a swamp of sameness as they dramatically accelerate the quantity of their content production but not the quality. Their use of AI will be to scale quantity in the hopes that wins them the game, or at the very least frees up their time to do other things. And for some companies, that will be a win, and that’s okay. If your company blog is atrocious now, a completely machine-generated blog of mediocrity will be a GIANT upgrade for your company.

But what if you don’t want average? What if you aspire to more than mediocrity? What role do these tools play? Here’s the part everyone is overlooking: these tools are better at refining than creating, and that’s the secret we need to understand to unlock their power.

Because these models – their technical name really is transformers – are adept taking in inputs and transforming them into outputs, they are actually BETTER at refining text than they are creating it. About a third of this article, the first third, was written with the help of ChatGPT. But it’s not what you think – it took my words and just cleaned them up. Here’s how – I did a voice recording while I was waiting to pick my kid up from art class, and fed it to Otter:

Otter transcript

then took that transcript and fed it to ChatGPT:

ChatGPT input

Is what you’re reading my words? Yes. It’s my words, but changed from one medium to another and cleaned up. My words were transformed by the GPT model – which stands for generative pretrained transformer – into text that’s almost exactly what I said, minus some things that weren’t helpful.

This is what these tools excel at – taking data and transforming it, rearranging it, making it more useful. This preserves our originality, our ideas, our language, while improving the quality – and that’s what they’re best at. Because they’re not relying on a gigantic average of all the content they’ve ingested, because they’re using our own words and just cleaning up or rephrasing, they perform great AND keep the spirit of what we’re trying to say. There was a great story on Buzzfeed about an AI app made for a contractor who is dyslexic, helping refine the inputs into better quality outputs.

We can even use multiple, different voices to create something useful from original inputs. My martial arts teacher, Mark Davis, has said it’s challenging sometimes to create social media copy that resonates with audiences. What better way to create ads than to use the voice of the customer itself? I wrote this prompt for GPT-3, using real customer reviews from the school’s Google Business profile:

Martial arts school GPT-3 prompt

What will happen? The large language model will digest not just my directions, but also the language of what customers had to say in their reviews of the school, then generate social media copy based on that. It’ll preserve the main ideas, the original ideas it was provided rather than resort to dipping into the pool of average commentary about martial arts schools.

And what was the outcome?

GPT synthesized reviews

Those are some good suggestions to get started with social media content. They’re clear, they’re specific, and they’re appealing.

This is the power of transformer-based large language models. You can have them create something average from scratch, or provide them with the raw materials and they’ll create refined products – but keep your originality and spark in the final product.

Now you know the secret!

Got a Question? Hit Reply

I do actually read the replies.

Share With a Friend or Colleague

If you enjoy this newsletter and want to share it with a friend/colleague, please do. Send this URL to your friend/colleague:

ICYMI: In Case You Missed it

Besides the newly-refreshed Google Analytics 4 course I’m relentlessly promoting (sorry not sorry), I would recommend the livestream we did this week on customer lifetime value. It really illustrates just how complex this seemingly simple calculation can be.

Skill Up With Classes

These are just a few of the classes I have available over at the Trust Insights website that you can take.



Get Back to Work

Folks who post jobs in the free Analytics for Marketers Slack community may have those jobs shared here, too. If you’re looking for work, check out these five most recent open positions, and check out the Slack group for the comprehensive list.

Free Book! Almost Timely 2022 Essays

I took all the letter parts of every Almost Timely newsletter from 2022 and put them together as a book. To my great surprise, it weighed in at almost 50,000 words, which is the average length of a business book these days.

However, instead of the usual price or filling out a form, I’m just giving it away, no strings attached. You can download it here in three formats, no registration or anything needed:

👉 Click here to download 📘 in PDF format

👉 Click here to download 📙 in Mobi format for Kindle

👉 Click here to download 📕 in EPUB format for other book readers

What I’m Reading: Your Stuff

Let’s look at the most interesting content from around the web on topics you care about, some of which you might have even written.

Social Media Marketing

Media and Content

SEO, Google, and Paid Media

Advertisement: Google Analytics 4 for Marketers (UPDATED)

I heard you loud and clear. On Slack, in surveys, at events, you’ve said you want one thing more than anything else: Google Analytics 4 training. I heard you, and I’ve got you covered. The new Trust Insights Google Analytics 4 For Marketers Course is the comprehensive training solution that will get you up to speed thoroughly in Google Analytics 4.

What makes this different than other training courses?

  • You’ll learn how Google Tag Manager and Google Data Studio form the essential companion pieces to Google Analytics 4, and how to use them all together
  • You’ll learn how marketers specifically should use Google Analytics 4, including the new Explore Hub with real world applications and use cases
  • You’ll learn how to determine if a migration was done correctly, and especially what things are likely to go wrong
  • You’ll even learn how to hire (or be hired) for Google Analytics 4 talent specifically, not just general Google Analytics
  • And finally, you’ll learn how to rearrange Google Analytics 4’s menus to be a lot more sensible because that bothers everyone

With more than 5 hours of content across 17 lessons, plus templates, spreadsheets, transcripts, and certificates of completion, you’ll master Google Analytics 4 in ways no other course can teach you.

If you already signed up for this course in the past, Chapter 8 on Google Analytics 4 configuration was JUST refreshed, so be sure to sign back in and take Chapter 8 again!

👉 Click/tap here to enroll today »

Tools, Machine Learning, and AI

Analytics, Stats, and Data Science

All Things IBM

Dealer’s Choice : Random Stuff

Advertisement: Ukraine 🇺🇦 Humanitarian Fund

If you’d like to support humanitarian efforts in Ukraine, the Ukrainian government has set up a special portal, United24, to help make contributing easy. The effort to free Ukraine from Russia’s illegal invasion needs our ongoing support.

👉 Donate today to the Ukraine Humanitarian Relief Fund »

How to Stay in Touch

Let’s make sure we’re connected in the places it suits you best. Here’s where you can find different content:

Events I’ll Be At

Here’s where I’m speaking and attending. Say hi if you’re at an event also:

  • Martechopia, London, March 2023
  • B2B Ignite, Chicago, May 2023

Events marked with a physical location may become virtual if conditions and safety warrant it.

If you’re an event organizer, let me help your event shine. Visit my speaking page for more details.

Can’t be at an event? Stop by my private Slack group instead, Analytics for Marketers.

Required Disclosures

Events with links have purchased sponsorships in this newsletter and as a result, I receive direct financial compensation for promoting them.

Advertisements in this newsletter have paid to be promoted, and as a result, I receive direct financial compensation for promoting them.

My company, Trust Insights, maintains business partnerships with companies including, but not limited to, IBM, Cisco Systems, Amazon, Talkwalker, MarketingProfs, MarketMuse, Agorapulse, Hubspot, Informa, Demandbase, The Marketing AI Institute, and others. While links shared from partners are not explicit endorsements, nor do they directly financially benefit Trust Insights, a commercial relationship exists for which Trust Insights may receive indirect financial benefit, and thus I may receive indirect financial benefit from them as well.

Thank You

Thanks for subscribing and reading this far. I appreciate it. As always, thank you for your support, your attention, and your kindness.

See you next week,

Christopher S. Penn

You might also enjoy:

Want to read more like this from Christopher Penn? Get updates here:

subscribe to my newsletter here

AI for Marketers Book
Take my Generative AI for Marketers course!

Analytics for Marketers Discussion Group
Join my Analytics for Marketers Slack Group!


4 responses to “Almost Timely News: What ChatGPT is Really Good At, Measurement Strategies for Agencies Course”

  1. […] Almost Timely News: What ChatGPT is Really Good At, Measurement Strategies for Agencies Course […]

  2. […] Almost Timely News: What ChatGPT is Really Good At, Measurement Strategies for Agencies Course […]

  3. […] Almost Timely News: What ChatGPT is Really Good At, Measurement Strategies for Agencies Course […]

  4. […] Almost Timely News: What ChatGPT is Really Good At, Measurement Strategies for Agencies Course […]

Leave a Reply

Your email address will not be published. Required fields are marked *

Pin It on Pinterest

Share This