Some Personal News

#31 Using insights from one channel to hyper-optimize others and AI's impact on donor experience (and your organization)

Mar 19, 2023

Happy Sunday. A very warm welcome to all the new subscribers. I’m thrilled to have you as readers and truly appreciate your feedback and support.

In this post I’ll cover:

How AI will help you evolve your content, chat, research and data.
Using insights from one channel to hyper-start (and hyper-optimize) others.
What I learnt from a copywriter this week.

Let’s dig in.

Operations to Evolve in 2023

A lot of AI tools are fun but they’re not actionable. This week we all got new superpowers. GPT-4 launched. It was eye-popping.

Here are a some areas where I’m using recent developments in AI to improve donor experience. Why not consider following suit?

1. (Localize) Content: AI tools will continue to make things like creating, re-purposing, and localizing content easier. Localizing is under-appreciated. It will allow orgs to market internationally more quickly.

AI will continue to commoditize content, forcing us to get much better at adding a unique perspective to the things we publish.

2. Chat: I'm so bullish on this right now. You can now run chat across your entire donor journey 24/7. You'll always have someone on chat answering questions, managing donor relations and educating prospects.

AI will make chat a ubiquitous part of the donor experience.

3. Research: I’ve mentioned before how AI chat will cannibalize search. But, there will also be powerful research assistants for specific segments of information like Elicit.org, making it easy to ask questions about the efficacy of your programs and outreach.

4. Data: How many BI reports have you looked at this week? How much of your life have you spent waiting for them to load? Why not feed all your data into an AI assistant who you can then ask questions in natural language about that data?

Other areas include video, creative and website design. Did you read about someone this week drawing an idea for a website on a piece of paper, snapping a photo of it and requesting the AI to convert what it analyzed in the image into HTML? The AI replied almost instantly, converting what was hand drawn on the paper into a working website.

Converting what you’ve drawn in real-life into working code for a variety of applications such as landing pages or data analysis will be table stakes soon enough. Taking campaigns from conception to market in less than a day, spotting and analyzing trends in program finances or penning the annual report, are all becoming more achievable and inexpensive. I hope what follows is a positive correlation with AI adoption and the efficiency with which we’re running our organizations. It’s there for the taking.

How to Leverage Insights from One Channel to Hyper-Start (and Hyper-Optimize) Another

I shared the following image as a “first step to building a mini-CDP” before:

Text, table

Description automatically generated

It starts with identifying all the steps donors can take with your organization – and selecting the steps and journeys that are important versus the ones that can be ignored. And then there are some basics which are almost “universally true” for every organization aiming to secure their first wins and get a consistent, break-even ROI. Hopefully some of the Google Ads tactics I’ve been sharing have already brought you there.

The industry seems to be talking a lot about channels at the moment and what’s happening within them – new formats of video, P Max, new AI technologies, new social networks, retail media networks, creative variations, new-new-new. And figuring out the most effective way to reach new potential donors is another part of a CDP’s functionality (customer data platform) that every Non-Profit can implement without over-investing in technology.

The audience must always come first but channels are the levers to pull, and those levers constantly change. The “One Report” approach I shared in SPN edition #29 helps you find a place for new channels – efficiently – while still putting the audience first.

Keeping 10-15% of your budget aside for testing should make sense for most orgs. I suspect that with growing and gaining efficiencies from your core tactics, you’re already saving more than 15% with improved performance. That said, leveraging existing information from other channels is vital.

Let’s take a practical example and tackle YouTube Shorts. How can an organization maximize its chances of success from the get-go? Consider the following:

Identify where it fits in the donor journey. Previously I shared that it’d likely be the best fit for Single-to-Monthly campaigns – mainly because of the creative type. Video content is compelling when it’s relevant. Still, short form has its limitations – it’s much easier to be relevant to those that already know about you without having the luxury of a longer explanation. So Single-to-Monthly is a safe bet here.
Identify what it’s most similar to from the current mix. YouTube Shorts is a video channel created to compete with TikTok. It’s a safe bet is to assume that YouTube Shorts will appeal to a younger slice of the regular YouTube audience but won’t draw much of the incremental audience – both amongst consumers and creators. It’s similar in Audience to YouTube, in Creative insights to YouTube and TikTok, and in functionality to Performance Max and Google Ads as the latest Google development.
What should be your initial setup? My “One Report” comes to mind! Below is a reminder of an imaginative campaign from our taxonomy exercise:

Points 1) and 2) above already helped me narrow down the scope of what I need to be looking for – all teams need is to create a setup for YouTube Shorts and to pull the report across all current campaigns with the following filters:

Table

Description automatically generated

You likely won’t have all of them available – but the more you have, the merrier. Pulling that report, teams can quickly identify the best-performing devices, targeting criteria, any necessary bid modifiers to select, and creative types (from your TikTok campaigns if those exist). For example, after doing this exercise you’ll likely launch the YouTube Shorts campaign with the following setup:

Targeting Type – 1^st Party Audience enriched with Google’s In-Market and Affinity, as those tend to perform the best within the Google environment.
Both genders, age group of 25-34 – older audiences are unlikely to perform well on TikTok and, therefore, YouTube Shorts.
Device – Mobile only.
Bid Strategy of Maximize Conversions since it usually performs the best for Google instead of Target CPA.
Bid Modifiers specific to your Google performance – maybe it’s an increase in a specific geo region or vice versa, excluding a particular In-Market audience that tends never to perform.

The goal of using previous campaign performance for the new launch is to maximize its chances of success – and not start over from scratch, especially while you have all this data already available (though it can sometimes be hidden). Cohesive taxonomy across all channels lets you quickly unhide it and reuse the insights you’ve already gathered.

Of course, the campaign you’d launch would still have plenty of room to optimize – and some of it will be unique to the platform. But returning to my original point, the above is what most CDPs do – they help you activate new channels based on similarities with the existing ones and audiences’ specifics. There’s a ton that can be achieved by getting creative with the basics.And you won’t be thousands of hours and millions of dollars in debt to big tech.

One final point: I’ve found this “relevance table” a helpful resource to keep in mind during some discussions. It helps to answer the “Where it fits into the journey?” and “What is it most similar to?” questions.

What I Learnt from a Copywriter this Week

Know the placement or traffic source.

Knowing the traffic source gives you an upper hand. If you know someone is coming from TikTok, where they’ve likely spent the last 25 minutes with their eyes wide open, flicking their thumb upward to get another dopamine hit, then tailor your copy or design to that. If you’re writing copy without the context of where the eyeballs come from, you’re driving in the dark with no headlights. The traffic source gives you context on the following:

What environment is someone in when consuming the copy?
How much time will someone likely spend reading it?
How punchy does the copy need to be? (think billboards vs. static Meta ads)
How hard do we need to work for a click?

Different ad placements should represent the brand or a particular program in different ways, making the person consuming it feel closer to the copy. Does that make sense? It feels a bit woo-woo, but good copy always elicits a reaction.

Write for skimmers.

When I ran Facebook ads for publishers in 2013, driving people to slideshows filled with banner ads (I apologize to those whom I sent to an advertising hell-hole), I obsessed over studying media companies crushing organic Facebook distribution like BuzzFeed, BoredPanda, and others. Here’s what their copy all had in common:

All copy is written with 5th-grade-level reading comprehension. No big words — everything was easy to understand so that you can read faster.
Anything worth noting or that you must read was bolded, underlined, or had a heading style attached to it. It was hard to miss.
Everything had a positive emotion by the end — in fact, the positive emotion was also why these sites got shared so much on Facebook.

Brands that write copy well do all this, but they also speak in benefits versus value props. You want to write copy so that, as you speak, someone can picture that scene in their head, with their donation at the center of it.

Ensure they take the scenic route.

There are always two ways to get somewhere — the boring and scenic routes. Growing up in the Cotswolds of England I learned young that it’s always worth taking the scenic route. And when it comes to copywriting, there’s no exception. Good copywriting, especially long-form copy, should take you down a journey, sounding different at different points.

If you can evoke a reaction from your copy (a “hell yeah!” or a “wow!!” or an “oh my goodness!”), then you’re doing great. If you can paint a scene in someone's head where they can almost daydream of their life with what your mission is executing on, then you’re doing fantastic.