156. SPN: Dark Social

The New Face of P2P Fundraising and Optimizing for AI models: search and content

Jun 08, 2025

A very warm welcome to all the new subscribers.

You’ve joined a community of over 3k marketing and fund raising operators at mission-driven Org’s. I’m thrilled to have you as readers and truly appreciate your feedback and support.

In this week’s SPN:

Search is no longer just about click-through rates, it’s about reference rates
AI models: what content to produce and where to distribute it
Creative and automation: Caveat emptor
Dark social is the new P2P fundraising (and attribution challenge 👀)
and, plenty of Jobs & Opps that took my fancy this week

Let’s jump in!

Donor trust starts with data privacy.
Whether your supporters are in the EU, Canada, or anywhere else, laws like GDPR, PIPEDA, and Quebec’s Law 25 define how we must collect, store, and manage donor data.
With Fundraise Up, you get:
Built-in tools to collect and manage donor consent
Controls to support data access, correction, and deletion requests
Secure data infrastructure that respects donor rights and expectations
When donors know their information is handled responsibly and securely, they’re more likely to give and give again.
Talk to Fundraise Up this week and understand how they can help you raise more money by exceeding your donor’s giving experience.
Game changer? It is for me!

Informational queries are more likely to trigger AI Overviews than transactional queries, for now.

Search

The genesis of digital advertising is search. It changed the game with signal - what donors were searching for - and certainty - the Google model knew if donors came and searched again, so we knew what worked and what didn’t.

Before search digital ads were just like trad ones - you bought the home page of the New York Times website just like you bought the front page of the New York Times paper.

But search also changed the content - building a site meant you had to embrace search engine optimization to maximize the search traffic.

As search developed into a multi $ billion business we’ve taken much of this for granted.

AI is changing how Google answers search queries. AI startups are driving significant traffic. TikTok, Pinterest and Reddit are increasingly driven by search. The App Stores are all about search - for mobile and for Streaming. Voice is (finally) getting used - especially by younger people.

More and more search is no longer just about click-through rates, it’s about reference rates: how often your Org or content is cited or used as a source in model-generated answers.

Donors, journalists, and policymakers increasingly turn to tools like ChatGPT for information. Visibility means being part of the answer, not just the search results. This shift is the poke I needed to lean more heavily into trying to measure for influence, authority, and brand awareness. To stay relevant, you need to optimize not just for search engines, but for what the models choose to reference.

Optimizing for what AI models choose to reference - especially in tools like ChatGPT, Perplexity, or Google’s AI Overviews - is about building semantic authority not than just search engine authority.

For Org’s, this means showing up not only in rankings, but in answers. The days of focusing on one keyword are over. Now we must understand the entire conversation around the keyword including intent and series of actions.

Some of this you’re already doing. Some of the work being done in SEO is absolutely feeding the models. What’s exciting me about this evolution in search is models are looking beyond keywords, focusing on concepts and relationships which create new ways to build brand awareness for LLMs.

This has implications for not only what content to produce (across text, image, and video), but also where Org’s should look to distribute messaging (website, media, expert, or community contexts).

Here’s what I’m testing, and some of the reasons behind why I’m doing it the way I’m doing it:

Authority -> Contextual Relevance

Old Approach: “We are a leader in X.”
SPN Approach: “Here’s how our work connects to the deeper needs of donors, communities, or systems.”

Action: Producing content that places the mission within broader narratives — e.g., “Why mobile clinics reduce maternal mortality in fragile states,” not just “Our maternal health program in Uganda.”

Models prioritize relationships between concepts, not brand slogans. So helping the model draw those connections by spelling them out is going to be valuable. Schema markup will become important because the engines will need to pinpoint the original source of the information serving up.

SEO Keywords -> Semantic Concepts

Old Approach: Focus on keywords like “disaster relief nonprofit” or “animal welfare charity.”
SPN Approach: Address the underlying intent and semantic relationships — e.g., “how to help displaced families after a hurricane” or “why pet fostering reduces shelter overcrowding.”

Action: Building content libraries that answer high-intent, real-world questions — not just what your org does, for instance, but why it matters in specific contexts.

Generic Content -> Use Case-Centered Messaging

Old Approach: “We provide food aid to children.”
SPN Approach: “In conflict zones, ready-to-eat packets prevent school dropouts caused by hunger.”

Action: Reframing impact stories and program descriptions around specific donor needs or scenarios. LLMs are going to elevate sources that explain the mechanism of impact, not just the output.

Link-Building -> Diverse Knowledge Contexts

Old Approach: Get backlinks from high-authority domains.
SPN Approach: Appear in diverse knowledge ecosystems - media articles, Reddit threads, YouTube explainers, expert interviews, PDFs, and community Q&As.

Action: Distribute content across multiple formats and platforms that AI models crawl. Think beyond your Org’s website.

Brand Visibility -> Model-First Discovery

Old Approach: Optimize for what Google ranks.
SPN Approach: Optimize for what language models remember and reuse - content that teaches, contextualizes, and feels trustworthy across time and format.

Action: Prioritizing evergreen, educational, and high-context materials. Don’t just tell donors what you do - tell models why what you do matters too.

A few weeks ago I mentioned sitting in the audience when Uncle Zuck shared his plans to fully automate ad creation.

He doubled down at the shareholder meeting last week and shared more detail:

Meta also plans to enable advertisers to personalize ads using AI, so that (donors) see different versions of the same ad in real time, based on factors such as geolocation. For example, someone in Florida might see a hurricane relief ad featuring local response efforts, while someone in California sees the same campaign focused on wildfire recovery - all driven by the same core message but contextually adapted to increase relevance and impact.

When I worked at UNICEF we were doing this sort of thing on Facebook ads - 4 years ago. Done properly, it’s highly effective but it won’t work if done badly and if overused.

Org’s need to remember that Meta has its own interests to protect. It has a huge volume of inventory to sell - good and bad.

Whilst your Org and agency want to maximize outcomes for your budget, Meta want to do that using the least amount of good inventory. So any secondary benefits, like branding for the 99% of people who don’t click, are minimized. Caveat emptor.

Jobs & Opps 🛠️

Cru: Digital Project Manager ($52,000)

Elton John AIDS Foundation: Chief Communications Officer

Girls Who Invest: Director, Individual Giving ($145,000 - $170,000)

International Rescue Committee: Officer, Strategic Events ($85,000 - $92,000)

UNICEF Denmark: Head, Global Partnerships

Global Fund for Women: Philanthropic Communications Specialist

Girls Who Code: Director, Individual Giving ($112,000 - $140,000)

Action Against Hunger USA: Chief Financial Officer ($335,000 - $350,000)

Prostate Cancer UK: Senior Strategic Communications (Campaigns) Officer (£34,300 - £37,300)

Save the Children: Managing Director, Philanthropy

→ Many more job opportunities listed on SPN’s sister site: Pledgr

Dark Social →The New Face of Peer-to-Peer Fundraising

Peer-to-peer fundraising isn’t disappearing - it’s just gone private.

Where supporters once posted donation pages publicly on Facebook, now they’re dropping links into group texts, WhatsApp threads, Slack channels, and DMs. This shift is part of what is called dark social - private sharing that’s nearly impossible to track, but incredibly effective at driving trusted action.

For nonprofit operators, it’s a dream and a dilemma: dark social delivers high-impact, low-cost reach - but breaks the attribution model. Traditional analytics don’t capture a WhatsApp forward or a message in a group chat. That doesn’t mean it’s not working. It just means you can’t see it.

According to RadiumOne, 69% of all social sharing happens in dark social. So this means rethinking how we measure influence and understand donor behavior.

What can Org’s do to track – and leverage – this “new” channel?

Here are 6 tactical recommendations split across Measurement - Activation - Quick Wins:

Measure - Unique Landing Pages

Let’s return to 2005 with this recommendation that has stood the test of time. Automate the creation of unique landing pages – strictly speaking, unique URLs – by geography and date, so that:

People in different geographies – ideally, cities – have different URLs, and
Every day, the URL gets modified, and
Every campaign – and more importantly, every influencer - receives a unique URL.
Finally, every different outlet – eg, a blog post posted on your Org’s Reddit page should have a different URL from the same post on Facebook.

Modifications can be a custom parameter at the end of the URL, but make sure it’s not a UTM parameter that some browsers might strip on a copy/paste action.

What matters is if that unique link is sent to a donor who clicks on it from the “wrong” location or on the “wrong” day, you can see that session in GA and trace it back to where the interaction originated from. You can also use tools like Unbounce or Instapage to automate the creation of new URLs based on multiple parameters.

SPN Tip: Technologically, this is usually done by using “hash fragments” – parts of the URL following a # symbol. Unlike UTM parameters, hash fragments are retained during copy-paste actions within messaging apps.

Measure - “Homepage Shadow Lift”

The hardest use case to measure is when supporters text or message direct links to one another to your homepage.

In GA4, this appears as “direct” traffic, with no further insight. One way to measure and isolate this behavior is to run what I’ve untechnically called a “Homepage Shadow Lift” analysis - comparing baseline direct traffic patterns and spikes aligned with known, “public” campaign pushes in order to identify the “shadow” content sharing peaks.

In Google Analytics 4, create a new custom report to analyze “direct” traffic segmented by hour and day.
Establish a clear baseline: Average daily direct traffic outside active campaign periods, increases in paid media budgets, or email sends.
Look for any inconsistencies. Did the percentage of direct traffic deviate by more than 10% WoW for a particular hour? Set a notification for that.
If the spike wasn’t associated with any known email pushes, influencer campaigns, etc., this might be the impact of “dark social sharing.”
Further segment the report by device and location to identify where the “dark social” might happen.

This isn’t a useful analysis to perform weekly or even monthly, but doing it once helps identify platforms on which to focus intentional attention for activation tactics.

Activate - Social Graph Enrichment

Tools like Clearbit - now recently part of the HubSpot family of brands - can add social profiles and influence metrics straight into the donor file. It can be used for mass donors - the cost isn’t prohibitive, and the process can be automated to a simple API pull once a week - but it is an overkill, and very hard to activate.

Where it becomes most useful is identifying major donors with sizable follower bases – and “sizeable” is anything above ~250 people. They aren’t influencers but active social media users. For all new donors added weekly above a certain $ threshold for your Org, automate the pull of their “friendship” connections with donors already on file.

Circles of “friends” will start emerging after the first few months. Use these circles to dig into persona traits of your Org’s new, younger major donors. Looking at them in groups will highlight patterns unnoticeable at the one-person level: tools they use (Discord? WhatsApp? Telegram?), topics they follow on messengers, groups they are part of on social media, etc. These then become targeting criteria for your paid media activities - and a great rolodex for the major donors fundraising team.

Activate - Messenger (and broader Meta) Advertising

Meta owns WhatsApp and their “intent” data already includes donors’ sharing content about your Org on WhatsApp - even though they won’t allow you to use it explicitly. Your Org is likely already using the Meta Ads and Advantage Plus. One extra thing to do is test Messenger advertising.

Create a Messenger ad campaign in Meta Ads Manager, include your best-performing creative, and use “Send Message” as the campaign objective. The algorithms will prioritize “dark social sharing” as the action to bid for.

Quick Win - “Share” vs “Send” buttons

Have you tested offering Facebook or LinkedIn’s “send” buttons on your content and not just “share”?

Implement them across your emails and relevant pages on your website, and add the “utm_source=dark_social” to size the opportunity.

Quick Win - Post-Donation Surveys

If your Org is already using them – add the “direct message from a friend/colleague” as an option above the fold in them.

I also like to track the “Shadow Shares” metric for each content piece that the Orgs I work with push out there - clicks from Messenger or WhatsApp, calculated either directly via UTMs, or indirectly via tactics #1 and #2 outlined above.

Over time, it’s helped me better understand topics that get redistributed. The result being I can adjust paid media creative on social, allowing us to capitalize on even a hint of virality.

The new face of peer-to-peer fundraising is quieter. It’s harder to track. But it might just be your Org’s most powerful channel yet.