From UNSPAM 2026 to action: 5 takeaways worth stealing 

There’s no conference like UNSPAM. 200 (self-proclaimed) email geeks and 3 days of unscripted, unfiltered conversations with the people actually “driving DAUs,” “caring about DMARC and BIMI,” and sharing Claude use cases that actually do something useful in the world of email.

Three weeks after launching our AI Agent, built to help teams build, optimize, and iterate, I appreciated the timeliness of the moment: a chance to hear what lifecycle and growth marketers are facing, and what still needs to be solved. I attended 2 years ago in Greenville, SC, and found it to be one of the most actionable conferences I’d ever attended.

I’m emptying my notebook on the key takeaways, and what I’m doing about them.

Cyrill Gross, who runs Mailix, gave a talk that emphasized the technical keys to accessibility, and why it’s not just a “nice thing” to do, but truly an email performance booster.

Megan Boshuyzen, email legend and developer at Inbox Army, called out during the ESP session how drag and drop editors and AI are not giving marketers the tools needed to optimize email code for accessibility.

The problem I see here—while many teams have good intentions, they don’t have the expertise or technical access to improve their accessibility through raw code changes, and ESPs make campaign-building simpler for the masses using drag-and-drop and no-code features. The question—how can we as ESPs help solve this?

We launched the Customer.io AI Agent to help lifecycle teams remove technical blockers and automate tedious tasks. I’m going to work on building an “Accessibility Audit” workflow and share it for others to use.

Maggie Glascott, Senior Lifecycle Marketing Manager at Buffer, flagged a past campaign example that was edgy and drove engagement, but also caused a lot of frustrated replies.

In the same vein, Hannah Castellanos, Marketing Auto Campaign Manager at TA Instruments talked about how lifecycle marketers need to gut check your campaigns to make sure you’re not "babying" your audience, when in fact, they are more than likely experts in your niche. An example: if your audience is a group of scientists, speak to them as a peer. You aren’t their teacher and they’ll immediately unsubscribe from any content that puts them in the position of “student” when really it’s the other way around.

Hannah outlined a 4 step framework that she uses for decision making in campaigns to make sure the advantages of the experiment are worth the trade offs. I'm going to use this for auditing content and copy:

Ask: Is my content...

1. Specific? (not vague ramblings that reveals the writer doesn’t know about the topic)
2. Does it acknowledge the uniqueness of the topic?
3. Are you speaking to them as a peer?
4. Is your CTA aligned to actions they can realistically take? (example — if a buying cycle takes 1 year, sending a “buy this week” offer erodes trust)

I'm also going to use Maggie's question to avoid launching campaigns that improperly prioritize performance over the risk of a good customer experience:

“If this experiment only raised engagement by .5% would it be worth the trade off? (Losing trust with some)”

My Grandma and I made a mistake our first time in Spain. After a long day of travel, we were famished and popped into a tapas bar. Each serving was small, so we loaded up, grabbing 5 plates each and heading back to our table. Aside from getting dirty looks, our table was overwhelmingly full, and it was a chaotic meal. I don’t remember a single dish we ate.

What a shame. Here we were in front of all this great, interesting food meant to be eaten one by one, and we couldn’t even process or enjoy the meal as it was meant to be enjoyed.

Kelsey Yen, Lifecycle Marketing Manager at Beefree, made this analogy beautifully in her talk, emphasizing how your content should bring content to your customers in “courses” not dropping 10 plates on their table at once and saying “good luck digesting all of that!”

Ethan Norville, Email Marketing Manager at The American Kennel Club, gave a talk that reinforced this concept in his talk on the “Messaging Gap.” Be sure to use 3 ingredients in an email to keep it aligned to an easy to understand, and valuable:

• Emotional cues: are you using imagery in your hero to build emotion around your message?
• Purpose: If theres not a clear “why” we’re sending this message, don’t send it.
• Usefulness: Do you understand your subscribers needs well, and are you delivering something of value to their need at this specific moment?

I’m going to subscribe Claude to our emails and ask it to grade each email on these 3 ingredients and Kelsey’s “eat in courses” analogy.

Kasey Luck talked about the problem with marketers and AI right now: "You hired AI to be your intern, but you’re asking it to be your CMO.”

She explained that you should lean on AI to process manual tasks involving large data or content sets, but avoid asking it to make judgments and strategy decisions. That’s where you, the human marketer, know best. (That's what we believe at Customer.io, too).

She shared an actionable playbook for how you can use Claude to glean audience insights and launch related campaigns more effectively.

• First, give Claude a CSV export of your brand’s reviews, forum discussions, customer feedback or sales call transcripts.
• Ask Claude to categorize the reviews into common categories, but also include 10 required/transcripts verbatim for each - so you can both find insights and scale, and understand the true way a customer talks a about it.
• Key - ask Claude to back up these categories and takeaways with multiple pieces of evidence, so you avoid it overemphasizing a “one-off”.
• Review Claude’s output and insights and then use it to audit underperforming parts of your funnel or lifecycle and use your findings - are you speaking to each persona/category with the right messaging?

I’m going to test out Kasey's playbook myself. I want to see how our messaging changes when we align it to personas and launch related campaigns.

A long, long time ago, in the 2010s, marketing tools synced data nightly or even weekly in some cases.

Real-time data in marketing systems and triggered campaigns changed the game, and soon this was table-stakes. AI became mainstream a few years back and increased the importance for good, accurate data. And while we’ve all made strides, in UNSPAM round tables, lifecycle marketers flagged that data accuracy and timeliness are still major issues and not solved universally yet.

One marketer flagged that Ecomm purchase data takes 30 mins to sync from Shopify to their email tool, meaning that their abandon cart email frequently triggers to consumers who actually already made the purchase in that lag time.

A B2B marketer flagged that relational data is still hard to use in marketing systems. Most folks are used to the contacts <> accounts data relationship that many CRMs like Salesforce are built around, however, many ESPs and marketing automation platforms still struggle to accomodate this type of data in a way that allows marketers to segment, personalize, and feel confident they are messaging all contacts at an account with the right context.

People try to hack this by mapping account attributes to a contact profile, but that quickly leads to data mess and confusion. Imagine a company is headquartered in Dallas. Claire from Portland has a “Dallas” attribute mapped to her contact profile because she’s related to her company’s account object. When pulling a list of “Dallas leads”, Claire gets invited to an event she is 2,000 miles away from.

Marketers flagged that former industry leading tools, like Marketo, Eloqua, and Salesforce Account Engage (Pardot), are engrained in their tech stack, but still keep a lot of things locked in code or inaccessible for their teams. They need flexibility and agility, but there’s often still friction trying to launch campaigns with these tools.

We’re a B2B company ourselves, and we’ve been able to solve for many of these data issues. I’m going to record a Loom sharing how we keep data clean, timely, and actionable for our team.

The conversations happening UNSPAM—about accessibility, trust, data, and how AI fits into all of it—are the ones that shape where email goes next. I'm taking these five things back to the team and putting them to work. If any of these ideas sparked something for you, I'd love to hear what you're doing about it. Reach out to me on LinkedIn, or at cody@customer.io.