Cloudflare and beehiiv give publishers new AI crawler controls

/0 Comments/in , /by
AI beehive crawlers

Cloudflare and beehiiv added AI crawl controls to beehiiv’s platform. This gives newsletter publishers a way to see, allow, or block AI bots from their dashboard as AI search becomes a new discovery path for web content.

The integration, announced Tuesday, embeds Cloudflare’s Crawl Control technology into beehiiv. It lets publishers manage how AI search engines and agents access their content, either by allowing crawlers for broader discovery or blocking scraping to protect archives for future licensing and monetization.

AI bot data comes to the dashboard. beehiiv publishers will get an on-platform dashboard showing which AI crawlers tried to access their content, which were blocked, and how much referral traffic those crawlers sent back. The dashboard gives publishers a side-by-side view of crawler activity, blocking decisions and referral traffic from AI services.

Publishers get simpler controls. The companies said publishers will be able to allow or block specific AI models with one-click permissions. Cloudflare will also update the system as new AI crawlers appear, reducing the need for publishers to manage robots.txt files, firewalls, or code changes themselves.

What they’re saying. Cloudflare CEO Matthew Prince said the partnership gives newsletter operators “transparency and control” as the internet changes; beehiiv CEO Tyler Denk said publishers need “real leverage” as AI changes how people find and consume content. From Cloudflare’s announcement:

  • “As AI models evolve to offer new forms of search and discovery, independent creators are looking for flexible ways to understand and manage how their content is accessed. This integration simplifies the process by letting beehiiv users manage their digital footprint through two clear choices: publishers can either opt-in to maximum discovery to allow AI search engines and agents to crawl their work freely for broader distribution, or choose content protection, blocking AI scraping to preserve their archives for future monetization and licensing opportunities.”

Why we care. The key question is whether publishers will actually use these controls once they are available. AI crawling has outpaced many creators’ ability to manage it, and adoption will show whether simple dashboard controls are enough to change publisher behavior.

Rollout starts now. The new controls are rolling out through beehiiv’s standard dashboard settings. All beehiiv users will get beta access to AI Crawl Control for visibility into AI crawler activity and traffic. beehiiv Max customers will also be able to block AI crawlers.

The announcement. Cloudflare and beehiiv Introduce AI Crawl Controls to Help Independent Publishers Navigate the AI Era

Read more at Read More

Shopify launches AI-powered marketing automation tool

/0 Comments/in , /by

Shopify is introducing Campaign Autopilot, a new AI-powered marketing tool that automatically creates, manages and optimizes campaigns across multiple channels, reducing the need for merchants to manually manage advertising and email marketing.

The feature is launching in early access and is available directly within the Shopify admin.

What’s happening. Campaign Autopilot uses AI to plan and run marketing campaigns on behalf of merchants across channels including Meta, Shop Campaigns and email.

Additional channels are already on the roadmap, including ChatGPT Ads, Microsoft Advertising and Snapchat.

Rather than requiring merchants to build campaigns manually, the system handles campaign creation, budget allocation and ongoing optimization automatically.

Why we care. Campaign Autopilot lowers the barrier to running multi-channel marketing campaigns by automating much of the work traditionally handled by agencies or in-house specialists. Instead of managing separate campaigns across Meta, email and other channels, merchants can set a budget and goals while Shopify handles campaign creation, optimization and budget allocation.

How it works. Merchants set a monthly budget, choose which channels to connect and define approval rules and guardrails.

From there, Campaign Autopilot:

  • Creates and launches campaigns.
  • Allocates budget across channels.
  • Adjusts spending based on performance.
  • Recommends email automations.
  • Monitors results and makes ongoing optimizations.

Merchants can approve campaigns before launch, modify budgets or pause activity at any time.

What’s different. Shopify is positioning Campaign Autopilot as an alternative to traditional campaign management tools and agency-led marketing.

The company says the system leverages performance data and patterns across millions of Shopify stores to inform recommendations and budget decisions.

Campaign Autopilot also operates separately from existing campaigns, meaning merchants already running Meta or Shop ads won’t see those campaigns altered.

The bigger picture. Shopify is increasingly embedding AI into merchant workflows, moving beyond ecommerce infrastructure and into growth and customer acquisition.

The launch reflects a broader industry trend toward autonomous marketing systems that can execute campaigns with limited human involvement while continuously optimizing performance.

What to watch. Shopify plans to expand channel support in the coming months, including integrations with ChatGPT Ads, Microsoft Advertising and Snapchat.

The company also says merchants can use its AI assistant, Sidekick, to review recommendations, trigger actions and monitor campaign performance.

Dig deeper. Introducing Campaign Autopilot: AI-powered Marketing Built into Shopify

First spotted. The update was spotted by Digital Marketing Consultant Susan Richards-Benson who suggested this feature for smaller ecommerce brands on Linkedin.

Read more at Read More

Straight from the source: 2026 Search Engine Land Awards judges reveal what makes an application award-worthy

/0 Comments/in , /by
search engine land awards winners
search engine land awards winners

Since its inception in 2015, the Search Engine Land Awards have recognized exceptional marketers on an annual basis — showcasing outstanding work, providing well-earned exposure in coverage and interviews, and bestowing upon them the highest honor in search.

While there’s no single formula for creating a winning entry, our judges have seen enough submissions over the years to know what separates the truly exceptional from the merely good. The strongest applications don’t just share results, they tell a story. They provide context, demonstrate strategic thinking, and clearly communicate why the work mattered.

And because great advice shouldn’t be gatekept, we thought we’d bring some of those insights directly to you.

We asked several of this year’s Search Engine Land Awards judges to share their best advice for prospective entrants. From common mistakes to avoid to the elements that consistently stand out, their insights offer a valuable look inside the judging process and can help you build a stronger, more compelling submission.

Keep reading for a roundup of fresh insights from some of our judges. (And see the complete list of 2026 judges here!)

“A great entry is a story with a goal, an action, and a measurable outcome that ties back to it. Tell that story as well as you can, and include a deck that makes it easy to see exactly what you accomplished.”

– Amy Hebdon, Founder, Paid Search Magic

“Explain your tactics. Many entries just say “we used best practices”. Everyone’s best practices and tactics differ. Explaining the process that lead to your results will highlight your creative thinking, problem solving, and uniqueness. Showing your insights and thought processes helps your entry standout and showcase your company’s competitive edge.”

– Brad Geddes, Co-Founder, Adalysis

Brad Geddes

“I look for SAY which stands for: Situation, Action and Yield.
Applicants should write a clear example of the situation, what they did and the result achieved over the time period,”

– Jo Juliana Turnbull, Growth Marketing Senior Manager, Holafly

“Show me the humans behind the metrics. We’re in a time where AI is reshaping search at a pace none of us have seen, and that shift matters…but the applications that rise to the top will lead with empathy, not just analytics. I want stories where I can see how your work built genuine trust with real people, not simply visibility in search and AI engines. I’m especially drawn to entries that embrace a wellness-based approach to their craft, and to teams who pair their quantitative wins with qualitative insight: the quote, the aha moment, the change in how someone felt about a brand or experience. Tell me how you held the human at the center – as strategy. If your project made people feel seen, understood, or genuinely helped, lean into that. Those are the stories I’ll be looking for.”

– Danita Smith, Founder & CEO/Chief Innovation Strategist, Adanis Design

“Clearly state the challenge you solved, and back it up with data. Explain the strategy behind the tactics you used and the results they drove. Tell me not just what happened, but what impact did it have on your campaigns? What did you do differently as a result?”

– Melissa Mackey, Director of Paid Search, Compound Growth Marketing

“Evidence: charts, analytics, screenshots. Be detailed, specific, and share data.”

– Barry Schwartz, Editor, Search Engine Land

“Tell a story. Numbers get you in the room, but the story is what stays with the judges. I want to know what the problem was, why it was hard, what you tried, and what finally worked. That arc, the messiness of real work, is what separates a memorable entry from a forgettable one. The submissions that stick with me are the ones where I can feel the thinking behind the decisions, not just the outcome. You did great work this year; now, make the judges feel the weight of what you solved before you show them the numbers.”

– Ameet Khabra, Founder, Hop Skip Media

“The two main things all award-winning entries share are that they explain the whys behind the hows, and they bring receipts (data to back up claims). If you can’t share the data behind your entry (budgets, revenue, etc.), you are putting yourself at a distinct disadvantage and may end up wasting the entry fee. A lot of people submit the same practices – if you can distinguish yourself by showing innovative thinking, you’ll do well!”

– Navah Hopkins, Product Liaison, Microsoft

“Give me all the data you can. Show the numbers and the real impact of whatever you did; conversions, ROI, and whatever monetary increases you were able to cause.”

– Celeste Gonzalez, Content Implementation and Product Specialist, Lastmile Retail

“Show me something I haven’t seen before, then prove it worked. The applications that land are the ones with a genuinely unexpected approach backed by numbers that make the result undeniable.”

– Adam Tanguay, Head of Growth, Jordan Digital Marketing

“I am looking for an approach or strategy that challenges the norm of SEM. A unique approach that focuses on achieving the business goals by way of campaign structure across Brand, Non-Brand, Performance Max, Conquesting, and general upper-funnel tactics. An advanced way of thinking about the target audience, messaging, conversion goals, etc. that helps show a sophisticated way of managing the campaigns & overall strategy to exceed business goals.”

– Matt Devinney, Director, Client Partner, Tinuiti

“I am looking for projects that break new ground with innovative takes on SEO, and are backed up by data and numbers-driven insights every step of the way.”

– Olya Ianovskaia, Founder and Lead Consultant, MycoMinds SEO

“Make your entry easily readable. We are going to need to go through several entries – I know the entries could be quite technical (and the quality of that will take precedent), but I am more likely to vote for you if I enjoyed reading your entry.”

– Anu Adegbola, Paid Media Editor, Search Engine Land

“My #1 piece of advice is to showcase strategy that truly breaks new ground. Award-winning applications demonstrate innovation that anticipates where SEM is heading, whether that’s leveraging AI in novel ways, pioneering audience-targeting approaches, or developing unique cross-channel integration. But innovation alone isn’t enough. The most compelling entries connect these forward-thinking strategies directly to measurable business outcomes, providing clear evidence of how your work translated to client growth metrics that matter. We’re looking for that perfect balance: creative execution that pushes boundaries while delivering documented ROI that proves your approach wasn’t just innovative—it was transformative.”

– Joseph Kerschbaum, Senior Vice President, Search & Growth Labs, DEPT

And there you have it! Submit your entry today to be considered by this year’s esteemed judges. Early Bird rates expire July 10… so get a move on!

Read more at Read More

YouTube rolls out new Gemini-powered insights tools

/0 Comments/in , /by

YouTube is expanding its suite of creator marketing and campaign intelligence tools with new Gemini-powered features designed to help brands identify trends, understand creator audiences and improve campaign performance.

What’s happening. Google is introducing several new insights and optimization tools across YouTube and Google Ads that give marketers more visibility into trends, creator performance and audience behavior.

The company says the new capabilities are intended to help advertisers make better creative and media planning decisions in an increasingly AI-driven marketing landscape.

Why we care. These updates provide deeper visibility into what’s trending on YouTube, which creators are resonating with audiences, and how their brand is performing across both paid and organic content. That can help marketers make smarter decisions about creator partnerships, campaign planning and creative strategy.

What’s new:

More detailed trend insights.

Google Ads’ Insights Finder is gaining expanded trending insights in the U.S., providing advertisers with a more granular view of what’s gaining traction on YouTube.

Brand Pulse data comes to Insights Finder.

Select Brand Pulse metrics are now being integrated into Insights Finder, allowing brands to evaluate both their paid and organic presence in a single location.

New creator insights API.

The new Content & Creator Insights API gives agencies and partners deeper information about YouTube creators and their audiences, helping improve media planning and creator selection.

Gemini-powered creative recommendations.

Google says Gemini will soon provide creative optimization tips for Demand Gen campaigns, including recommendations on visuals and creative elements that may improve performance.

The bigger picture. As creator-led content becomes increasingly influential in purchase decisions and brand discovery, advertisers are looking for better ways to identify emerging trends and measure creator impact.

Google is betting that AI can help marketers surface those insights faster and make campaign planning more efficient.

Bottom line. YouTube is giving brands and agencies more data on trends, creators and campaign performance, while using Gemini to help turn those insights into stronger creative and media decisions.

Read more at Read More

How to make Performance Max focus on net new customers

/0 Comments/in , /by
How to make Performance Max focus on net new customers

There’s a trap door waiting for DTC brands that invest in Google Ads that makes your dashboards look amazing, but absolutely wrecks your P&L.

It’s the danger of recycling traffic from Meta.

Thanks to the overlap between paid search and paid social traffic, running Google as a standalone channel is incredibly difficult if you don’t know how to set it up. Ad platforms refuse to share data with one another, and they love to claim credit for the same conversion — even if those sales would’ve happened without the influence of ads.

The DTC brands I speak to are often proud to show off their new customer numbers: month-over-month growth, a steady upward trend, and a fantastic dashboard. But when we go deeper into the data, we often find that a big chunk of those “new” customers are:

  • Conversions that would’ve happened because of brand or content efforts.
  • Customers who aren’t truly incremental because they consumed ads on multiple platforms.
  • The same people signing up with multiple email addresses.

You could argue that these overlapping sales still count as revenue, and they do. But when you look at the contribution margin from those sales, they cost far more than they should and erode actual profit.

In other words, you lose money when you run ads on both platforms without guardrails.

But that doesn’t mean you need to stop or limit yourself to one channel. Instead, you need a better system for measuring actual customer acquisition.

Why the new exclusions matter

If you’re spending five figures or more on Meta, TikTok, AppLovin, or any other top-of-funnel channel, you’ll want to minimize overlap with other channels to drive actual new customer acquisition.

Here’s what that looks like:

  • Someone sees your ad on Facebook or Instagram.
  • They visit your site, browse, and leave without buying.
  • A while later, they search for your brand on Google or get retargeted on YouTube.
  • Performance Max swoops in, grabs the conversion, and reports strong ROAS.
  • You may have won that order anyway, but now Google and Meta both want credit for it.

Now you’re paying two or more platforms to recycle a conversion that you might have earned with just one.

Ever since Performance Max launched, there wasn’t much you could do about this. It’s been a bit of a black box that automatically goes after the warmest traffic it can find: branded search, site visits, email subscriptions, and existing customers.

It lets you bid more for new customers, but you can’t really stop the campaign from defaulting to easy mode.

A while ago, Google began letting you exclude people searching for your brand on Search and Shopping. Performance Max still targeted warm audiences through YouTube, Gmail, and the Display Network.

The latest round of updates from Google has finally addressed this problem. You can now force Performance Max to focus on net new customer acquisition through a combination of brand exclusions, audience exclusions, and Customer Match data. 

First-party audience exclusions, announced in March, are the final piece that makes this possible (though not foolproof – customer list matching is never perfect).

See exactly how your competitors win.

Uncover the keywords, ads, landing pages, and strategies driving your competitors’ paid search success—and find your next opportunity to outperform them.

Analyze your competitors

A four-step framework for net new customer acquisition

Here’s a four-step framework we’re using at my agency to help clients maximize incrementality.

Step 1: Exclude your brand

This one has been around for a while, but it’s the foundation, so we have to start here.

For smaller brands, brand exclusions usually aren’t necessary. But once you’re spending real money and seeing more than 15% to 20% of your cost or revenue coming from brand searches, it’s time to take action. 

There are two parts to this.

Go into your campaign settings and add a brand exclusion. If your brand isn’t already on the list, click New brand list, create one, and add your brand. Google will do its best to block branded queries from this list.

Because brand exclusions aren’t foolproof, go to the Keywords tab inside the campaign and add your brand name as a phrase match negative keyword. Add a few common variations, too. This catches anything the brand list misses.

If you’re excluding brand terms from Performance Max, you need a dedicated brand Search campaign and a brand Shopping campaign to capture those searches. Otherwise, you’re just leaving money on the table for competitors.

Step 2: Exclude website visitors and email subscribers

Even if you blocked brand searches, Performance Max would still retarget people who visited your website, opened your emails, or interacted with your brand on YouTube, Gmail, Discover, and Display. So even with brand exclusions in place, a big chunk of your spend still went to warm traffic.

Now you can change that. Go to your campaign settings and find the new audience exclusions option. Then build a few remarketing lists:

  • All website visitors: Set this up through the Google Ads pixel or Google Analytics. It captures anyone who has visited your site.
  • Email subscribers: Connect Klaviyo (or whatever ESP you’re using) directly to Google Ads. The benefit of the Klaviyo integration is that the audience updates in real time, so new subscribers are added automatically.

Once you exclude these audiences, Performance Max can only go after people who haven’t interacted with your brand in any meaningful way. What we typically do, and what I recommend, is to come up with an engagement metric that fits each account’s business goal, such as cart adds rather than visitors from the past seven days.

What a change from how this campaign type used to work.

Get the newsletter search marketers rely on.

See terms.

Step 3: Exclude existing purchasers

Same idea as Step 2, but specifically for people who have already bought from you. You can do this two ways.

  • Through a pixel-based audience that captures anyone who has triggered the purchase event. 
  • By uploading your customer list directly. Shopify now lets you set up Customer Match lists right inside the Google Shopping app, and Klaviyo can do this, too.

Add these audiences to the exclusions section of your campaign, and you’re done.

A small caveat to keep in mind: audience matching is never 100%. If you upload a customer list of 1,000 people, Google might only match 900 of them. So you’ll still see some level of bleed. But going from “the campaign is targeting all my existing customers” to “the campaign is targeting maybe 10% of them” is still a huge win.

Step 4: Use ‘New Customer Bidding’ in campaign settings

The last piece is to tell the campaign explicitly that you want new customers.

In your campaign settings under customer acquisition, you’ll see two options: bid only for new customers, or bid higher for new customers. Both require you to connect a customer list (which you’ve probably already done by Step 3).

The “only new customers” option is the most aggressive setting. The campaign simply won’t bid on existing customers. Combined with the audience exclusions from Steps 2 and 3, this gets you as close to pure new customer acquisition as Performance Max will allow.

The “bid higher for new customers” option is more flexible. You set a dollar value that represents the additional value of a new customer, and the system bids more aggressively when it thinks an auction will result in one.

Here’s where you need to be careful. If you tell Google a new customer is worth an extra $100, and you get a $200 sale from a new customer, Google will report it as $300 in revenue. That extra $100 is a fictional reporting value, not real revenue. It will inflate your ROAS numbers and distort your target ROAS bidding.

Our recommendation is to use a small placeholder value, such as a penny or a dollar, when you want to nudge the system toward new customers without distorting your reporting. Or use a number that genuinely reflects the lifetime value premium of a new customer to your business.

What to expect from this approach

It’s still early, so we can’t draw firm conclusions yet. But based on my experience managing PPC for ecommerce brands, here’s what I expect to happen.

Many advertisers who walked away from Performance Max did so because it was simply recycling Meta traffic. By splitting it out, you force it to go after net new traffic.

This will likely benefit brands that don’t have a ton of video creative for YouTube, which is another platform where brands try to drive net new acquisition at the awareness stage.

One of the big differences between Performance Max and Demand Gen is that the former is much more conversion-focused. Any brand considering excluding branded Search and Shopping from Performance Max should also consider this tactic, as it tends to over-index on hot traffic.

In terms of outcomes, I expect the reported ROAS attributed to Performance Max to be lower than what you may have seen in the past.

But when you look at the breakdown of new versus returning customers, it should align much more closely with new customer acquisition. Without advanced configuration, it might be a 60/40 split, even in the best situations.

Limitations and realistic expectations

Nothing about this is foolproof. Audience exclusions don’t match perfectly. Brand exclusions don’t catch every variation. Customer Match has its gaps. So even with all four steps in place, some percentage of your spend will still hit warm audiences.

But for the first time, you actually have the levers to push Performance Max into upper-funnel territory. You can make it work like a real prospecting channel instead of a retargeting channel that takes credit for demand created elsewhere.

This matters most for brands spending heavily on Meta, TikTok, or other channels and wanting Google to actually grow the customer base rather than recycle the traffic those channels generate. If you’re seeing strong ROAS in Performance Max but flat new customer numbers month over month, this framework is for you.

If you’re a smaller brand still trying to find product-market fit or build initial momentum, this is probably overkill. Let Performance Max do its thing and pick up conversions without too many restrictions.

But once you’re scaling and the question is no longer “Can we be profitable?” but “Can we be profitable while growing the customer base?” these settings become some of the most important levers you have.

Every click they win is a customer you lose.

See where competitors are investing, which keywords drive their results, and how to capture more of the market.

See who’s stealing your traffic

Google’s giving you more control over PMax. Use it.

The conversation around brand versus non-brand is everywhere. You can’t throw a dart at a paid media conference without hitting someone with a strong opinion on it. But for some reason, almost no one seems to be testing this new option.

I just finished auditing an account spending $100,000 a month on Search with no Performance Max or Shopping, so they get purely new customer acquisition. We looked at their numbers and said maybe now’s the time to try this, exclude all these segments, and let it rip.

So here’s when I recommend implementing this test: if your ad spend is high enough (it doesn’t need to be $100,000 a month or anywhere near it), or you’re revisiting Performance Max. Your hypothesis should be that this approach increases the proportion of actual new customer conversions.

I think you’ll find that the needle moves further than you think.

Read more at Read More

How to approach build-versus-buy decisions for SEO

/0 Comments/in , /by
How to approach build-versus-buy decisions for SEO

AI has made SEO teams ambitious about what they can automate. Tasks that previously required engineering support can now be solved with the help of Claude or ChatGPT.

That’s exciting, but it also creates a new problem: thinking you can automate everything. In modern language, that often comes down to one question: Should we build or buy this new tool?

This build-versus-buy dilemma has never been simple, and AI has made it even more complicated. The challenge goes beyond cost. It involves security, maintenance, data access, internal capabilities, workflow fit, and whether a custom solution will remain maintainable, reliable, and useful six months from now.

How AI lowers the barrier to building

AI has lowered the barrier to experimentation. Even without technical knowledge, you can now create a custom GPT, build a workflow, connect data sources, or create an internal AI assistant.

But that doesn’t mean the same person can build and maintain a tool that will remain reliable over the next few years.

In most cases, AI can help SEO teams analyze data, identify patterns, summarize information, and recommend actions. It can save a lot of time, and teams that ignore AI are clearly falling behind.

But, at least for now, AI isn’t doing truly creative work in the same way humans do. It works from existing patterns and predicts likely outputs. That may change in the future.

AI also comes with hidden costs. Internally built tools are often treated as free because the invoice usually doesn’t sit with the SEO team. But that doesn’t mean token usage, API calls, infrastructure, engineering time, security reviews, and maintenance don’t cost money.

We are already seeing this effect. Reuters has described it as “corporate AI sticker shock,” with companies struggling to forecast usage-based AI costs. TechCrunch also reported that Uber introduced AI spending caps after blowing through its annual AI budget in four months.

Today, marketing teams aren’t the heaviest AI users, especially compared with engineering teams. But that can change quickly.

And when usage grows, the bills will grow too. That will naturally make companies ask which AI tools and AI-powered workflows create value and which ones only consume budget.

Be the brand AI recommends.

See where your brand appears in AI search, where competitors are winning, and what it takes to become the answer AI recommends.

See your AI visibility

Start by defining what you need

Before deciding whether to build or buy, SEO teams need to define what they really need.

Different ways to use AI and automation

Many teams group these solutions together, but they vary significantly in cost, complexity, and maintenance requirements.

  • A custom tool: A more complex internal system that usually needs engineering support. It is often more about automation, but it can have an artificial intelligence aspect.
  • A custom workflow: A repeatable process built with different tools, such as a custom GPT, Claude project, spreadsheet, reporting template, and so on. It often includes automation, for example, a scheduled task in an AI tool, and it usually has an artificial intelligence layer.
  • A custom layer on top of SaaS: Using data from existing tools and shaping it into your own reporting, prioritization, or recommendation workflow.
  • A true AI agent: A system that can take more autonomous actions. For example, it can scan your Slack and follow up with people you are still waiting on.

These aren’t the same, but people often label them incorrectly. Calling everything an “AI agent” creates confusion and can lead to wrong estimates about cost and complexity.

Look for repetitive, context-rich tasks

We’re still experimenting. Most of what our team has built focuses on daily tasks that require a lot of manual work.

For example, we’ve created a custom GPT that evaluates whether our content matches our personas and their pain points. The goal is not to replace the human copywriter or reviewer. It is to determine whether a piece remains generic and whether a few additions can make it more relevant.

We are also using AI for translations, monthly reporting, and a weekly summary that combines meeting notes, Slack, and Jira, and helps me see whether I have missed adding a task to Jira or where I still need to follow up.

One of our latest workflows transforms recorded internal meetings into organized landing page briefs.

These types of tasks are good candidates for AI-powered custom workflows because they rely on internal context, repeatable processes, and company-specific knowledge.

Get the newsletter search marketers rely on.

See terms.

Not everything should be built

One example from our team was a prompt tracking tool that my colleague vibe-coded. It worked well as a starting point. But the data presentation was not perfect, and it was hard to create a trend graph without additional manual steps.

Soon, it became a maintenance burden because every external change in any of the LLM tools required fixes, for which we needed engineering help.

The real issue was reliability. For AI visibility and prompt tracking, we needed consistent data in one place, presented in a way we could analyze over time. That is why we moved to a specialized platform like Peec AI instead of continuing to maintain our own version.

That experiment was still valuable. It helped us understand the problem, the complexity, and the features we actually needed from an external vendor.

And this is one of my pieces of advice: whether you want to build a tool internally or buy one, always test what is already available on the market. Only then will you really understand what you actually need. You may think you need 10 features, only to realize you use only three.

For business-critical tools such as rank and AI visibility tracking, and website crawling, small SEO teams without dedicated technical support should usually be careful about building from scratch. If the data is fundamental to decision-making, reliability should be your main decision factor.

Use AI where your data already lives

Buy the crawler, rank tracker, or AI visibility platform. Then focus your internal efforts on connecting data from these tools to custom information, such as your GA and GSC accounts or even CRM data. Once connected, create reports that combine all these sources and enable you to analyze everything in one place.

MCP connections are also worth considering. The Model Context Protocol is an open standard for connecting AI applications to external systems, data sources, tools, and workflows. With MCP servers, you can analyze data from your primary tools directly using AI, taking your current workflows to the next level.

This doesn’t mean you’re required to learn how to code. But they need to know enough to ask the right questions.

If a tool connects to an internal knowledge base, customer data, or proprietary research, you should be aware that this could pose a security risk. And it might turn out that it is better for the company to dedicate an engineer to support you rather than risk exposing sensitive information.

You should also understand what the final cost will be for your company when you decide to go with a custom tool. Custom tools aren’t free just because the invoice doesn’t sit with SEO. Engineering time, security reviews, AI tokens, and API usage are all part of the cost.

Before asking leadership for a tool, SEO teams should be able to explain the workflow problem, the expected value, the cost of buying compared with the estimated cost of building, and what might happen if nothing is done.

The best requests don’t start with: “We need this tool.”

They start with: “Here is the problem, here is why it matters, here is what we’ve tested, and here is the best way we think we can solve it.”

How to prioritize what to build first

There’s no single prioritization matrix that will work for every situation.

A website crawler, a content evaluation tool, a report builder, or a competitive intelligence system can’t be judged by the same criteria.

If you are in a situation where you think you need more than one tool, start by mapping your current workflow and what your ideal situation looks like.

Once you do that, the patterns will be clear. Often, your strongest priorities will fall into two groups.

The first are tools that can support revenue creation. SEO teams are usually part of the marketing organization, and marketing is expected to bring visibility or leads. If a tool can help identify content opportunities, improve conversion rates, increase AI visibility, or surface gaps versus competitors, it can be seen as a priority.

The second group is workflows and tools that can help you minimize repetitive manual work. This category may not create revenue, but it will give your team time back to focus on more strategic work.

Don’t forget that quick wins also matter. Stakeholders don’t want to wait three months before seeing results. A smaller project that can bring value in three weeks will help you build trust and make it easier to get support for bigger initiatives.

Cross-team value should also be part of your decision.

SEO problems are often not just problems for your team. Competitive intelligence, for example, matters to PPC, ABM, content, product marketing, and sales, too. If several teams share the same pain, the business case becomes stronger.

So don’t be afraid to act as a cross-team synchronization layer when needed. Talk to the same teams you have already worked with, and try to understand their workflows and pain points, and where your needs overlap.

And remember, the best tool is not always the most ambitious one. Starting with something small is often the smartest move.

If AI can’t find you, customers won’t either.

Track your visibility across AI search, uncover missed opportunities, and grow your presence where customers are asking questions.

See your AI visibility

Good decisions start with proper scoping

AI has made it easier to build, but that doesn’t mean you don’t need to think about what really needs to be built.

Before deciding whether to build, buy, or customize, take the time to properly scope the work.

  • Understand the problem, the value you expect, who will use the solution, and who will maintain it after launch.
  • Talk to your team and other teams. Determine whether this is only an SEO problem or a wider business problem.
  • Don’t build because AI makes it possible. Don’t buy because a demo looks impressive.

Without proper scoping, you can end up with an expensive SaaS tool that doesn’t fit your workflow or an internal tool your team can’t maintain.

Always think first. Dedicate enough time to scope properly. Then decide whether to build, buy, or customize.

Read more at Read More

Google Search Console AI performance reports rolling out to more users

/0 Comments/in , /by

Google has confirmed that it has expanded access to the new Google Search Console AI performance reports to more users. Google’s John Mueller wrote on Bluesky, “We’re just rolling these out incrementally to sites, and reviewing the feedback along the way. I know everyone wants the new shiny thing immediately… but first, patience.”

AI performance report. The report shows you how well your content and websites are performing in AI responses, AI Mode, and AI Overviews in Google Search. The reporting includes impressions, pages, countries, devices, and dates, but does not include click data. 

Expanding access. This morning, I spotted a number of SEOs posting that they are now seeing the report, and that it is not restricted to sites in the United Kingdom. Some are seeing the report for sites in the United States, India, Switzerland and so forth.

And as I quoted above, John Mueller from Google confirmed the search company is “rolling these out incrementally to sites.”

What it looks like. Here is a screenshot of this report:

Why we care. Site owners and publishers have been asking for controls over whether and how their content is shown in Google’s AI features since Google launched these a couple of years ago. Well, Google is rolling out this feature to more of its users today. It is unclear how soon everyone will gain access to these controls, but I am surprised that Google expanded access this quickly. Specifically within 20 days from its first release.

Read more at Read More

Why some channels reward breadth and others require commitment

/0 Comments/in , /by
Why some channels reward breadth and others require commitment

Many budget allocation strategies assume that every channel follows the same pattern: the first dollar is the most productive, and each additional dollar yields a slightly lower return.

The charts below show what that pattern looks like.

The log shape means that the first dollar is the most productive, and each subsequent dollar is worth a little less. When every channel looks like that, the game plan is to spread the budget to as many channels as possible and equalize the marginal CPAs to maximize profit.

But not every channel looks like that. Some have a warm-up region where the early spend is the least efficient, not the most. On those channels, the logic above breaks, and so does the “test small, scale the winners” playbook that most of the industry runs on autopilot. 

The difference comes down to one question about the channel: Is the response curve C-shaped or S-shaped?

The answer can change how you approach channel testing and channel measurement, including any MMM analysis. Moreover, Google has been incorporating more S-shaped campaign types, and after its Google Marketing Live announcements, this trend seems set to continue.

The two shapes — and the only part that matters

The response curve plots output (conversions, revenue) against input (spend). This generally results in two types of curves in marketing.

  • C-shaped (concave): Diminishing returns from the very first dollar. A log or power curve. Picture the top-left quarter of a circle: steep at the start, flattening as you go.
  • S-shaped (sigmoid): A slow, inefficient start, then an inflection point where it gets steep, followed by a flattening into saturation. A logistic curve.

The response curve itself isn’t what you allocate against. You allocate against the marginal curve, the derivative, which answers the question: “What did the next dollar buy me?” That’s where the shapes diverge in a way that matters.

  • For a C-curve, marginal return is highest at the first dollar and falls in only one direction. Marginal CPA rises from the first dollar onward. If conversions are a*ln(s), marginal conversions per dollar are a/s, so marginal CPA is s/a, climbing in a straight line as you scale. There’s no warm-up. The cheapest conversion you’ll ever buy is the first one.
  • For an S-curve, marginal return starts low, rises to a peak at the inflection point, then falls. Marginal CPA is U-shaped. It’s expensive at the start, bottoms out around the inflection point, then climbs into saturation.

That region of increasing marginal returns is the whole story. It’s the difference between a channel where small budgets are productive and one where they are wasted.

See exactly how your competitors win.

Uncover the keywords, ads, landing pages, and strategies driving your competitors’ paid search success—and find your next opportunity to outperform them.

Analyze your competitors

How this looks in a marketing campaign

Say your CPA goal is $50. Here is an S-shaped channel, modeled as Conversions = 1000 / (1 + e^(-0.25(s – 20))), with spend in the thousands and the inflection at $20,000/month:

Run the $10,000 test that a sane person runs before committing real budget. Average CPA comes back at $132, marginal around $94. If those two metrics are all you look at, you conclude that this channel can’t hit $50, so let’s kill it.

That verdict is wrong. At $20,000 to $25,000, the channel is running at an average of $32 to $40, and the marginal dollar in the $15,000 to $25,000 band costs $18. That’s not “barely viable.” In that band, it’s the best marginal buy you have. The small test fell within the warm-up and reversed the conclusion.

In a C-shaped channel, the small test would have shown you the best the channel can do. On an S-shaped channel, it shows you the worst.

This is the trap. The standard playbook is “test small, scale what works.” On S-curves, small tests systematically condemn channels that would’ve worked at scale because the test is structurally stuck in the inefficient region.

Get the newsletter search marketers rely on.

See terms.

The allocation logic, restated

C-shaped channels, go wide

The optimization is convex. There’s one global optimum, the equimarginal rule from the marginal-CPA post applies cleanly, and the solution is usually interior, meaning lots of channels get funded.

Even a small allocation is productive because the first dollar is the best dollar. Run many channels lean, reallocate continuously at the margin, and pull back the instant marginal CPA crosses your goal.

S-shaped channels, go deep or skip

The optimization is non-convex. A small allocation can be strictly worse than zero because below the inflection your marginal return sits under your target, and you’ve sunk money to get nowhere.

The decision isn’t “how much.” It is binary: commit past the threshold, or don’t fund it at all. There’s a real minimum viable budget, and it’s often above normal test budgets. You can’t sprinkle an S-curve and expect efficiency, and you can’t evaluate one on an underfunded test.

Those two rules can look like they fight each other, but that’s only true to a certain point. Past the inflection, an S-curve is concave, so the equimarginal rule governs it exactly as it governs a true C. The S-specific instruction — commit a block instead of sprinkling — is only about the trip from zero to past the inflection.

Shape is therefore mostly a launch-and-evaluation problem. Getting a new prospecting channel into its efficient range requires a committed block and patience with ugly early numbers. Once it clears the inflection, you manage it at the margin like everything else, right up until you consider cutting it hard, where shape matters again because the downside is a cliff, not a ramp.

This is the part that’s genuinely counterintuitive, and it echoes the original marginal-return point: The right move isn’t always the one that looks most efficient at a small scale.

Which channels are which?

The historical default was concave. Simon and Arndt reviewed more than 100 studies and concluded that advertising follows the law of diminishing returns, a concave response. 

The dissent came later: Vakratsas, Feinberg, Bass, and Kalyanaram found that threshold effects do exist and that response is not necessarily globally concave. Their explanation for why thresholds were so hard to find is the useful part. Mature accounts already operate inside the effective range, so the warm-up never shows up in the data, and most studies fit a concave model (the double-log) that can’t reject an S-curve even when one is present.

The platform shift has made the threshold visible again. Here is a fuller map, ordered roughly from C to S. The shape column is an inference from how each system targets and learns, not a measured constant, and the right shape for your account still has to be measured.

Two rows do most of the work.

AI Max is the live example of a channel migrating from C toward S. Swapping explicit keywords for broad and keywordless matching means it needs conversion volume to learn which queries convert, so below a data threshold, it explores badly.

The mixed independent results fit that: Google reports about 14% more conversions on average and up to 27% for exact-match-heavy campaigns, while independent testing reports 84% of advertisers seeing neutral or negative results. Much of that spread is accounts that turned it on without the conversion volume to clear the learning region.

Performance Max is the trap, because its curve is a composite. It blends a harvesting layer (branded, retargeting, Shopping against existing intent) with a prospecting layer (keywordless expansion across surfaces). The harvesting layer is a cheap C that pays off on the first dollar. The prospecting layer is the S underneath.

Blended, the early efficiency looks great, because you are mostly skimming demand you already had, and the average hides the prospecting warm-up entirely. That is also why the platform is glad to optimize it for you: the blend flatters the headline number. You can’t read PMax or run the shape analysis on it until you split the harvesting from the prospecting.

The throughline runs in two layers. Rules-based auctions capture the best inventory first, which yields concavity; machine-learning systems must be fed before they are efficient, which introduces a threshold. Underneath both, harvesting existing demand is concave and mostly non-incremental, while creating new demand is the S-shaped part where the real growth and the real warm-up cost both sit.

Average versus marginal: total over spend, or the slope where you stand.

What you allocate against is marginal incremental return, the slope of the incremental curve at your operating point. A holdout fixes the first axis only. Time-sliced marginal CPA on attributed data fixes the second only. A multi-cell scaling test gets both, at a cost. 

MMM (method 1) estimates the whole curve from aggregate data and sidesteps click attribution entirely, but pays in identifiability and modeling assumptions instead. Most arguments about ‘what is working’ are two people standing on different axes.

There are two major cautions, and I would flag both as genuinely unsettled rather than settled facts. 

  • Separating a true S-curve from “concave with a high half-saturation point” is hard, because a concave model will fit S-shaped data well enough to hide the inflection (this is the Vakratsas point, and it applies to your own dashboards as much as to academic studies). 
  • The learning phase may be a one-time fixed cost to train the model rather than a permanent feature of the steady-state curve. If it is transient, the channel may behave concavely at the margin once it is trained, and the S you measured was a startup artifact. The truth is probably a mix: a one-time training cost, plus an ongoing minimum-volume requirement to stay efficient. Treat every shape call as provisional and re-check it.

One more failure mode, and this one is not unsettled science but a matter of where you are standing on the curve. An S only looks like an S if your data spans the inflection. 

Above the inflection, an S is concave, mathematically identical to a C. Look at only the $20,000-and-up rows of the table above: marginal CPA rises monotonically from $18, a textbook C-curve, and the convex warm-up is invisible because you are no longer operating in it. 

Established accounts usually sit past the inflection, which is exactly why Vakratsas found thresholds so hard to detect, and why you can run an S-shaped channel for years, correctly, while believing it is concave. The tell arrives the day you cut hard and fall off the inflection instead of easing down a slope.

When to go wide and when to go deep

The marginal-return post told you to equalize marginal CPAs across the program. That rule is still correct, but the shape of the curve tells you how you’re allowed to get there. 

  • On C-shaped channels, you can get there by sprinkling, because every dollar is productive and breadth is the natural answer. 
  • On S-shaped channels, you have to commit a block of budget past the inflection before the channel earns its place, and then concentrate rather than spread.

Lay the harvest-versus-create cut on top. Harvesting channels (branded, retargeting, non-brand search) are your C-curves: fund the first dollars, then cap them early, because they saturate fast and most of the tail isn’t incremental, no matter how strong the attributed ROAS looks. 

Prospecting channels (Meta, YouTube, LinkedIn, the expansion half of PMax) are your S-curves and your only real source of incremental growth: commit past the warm-up or don’t start, and judge them on incremental lift rather than attributed CPA, or you’ll kill the thing that was working.

Classic search rewards going wide. PMax, AI Max, and Meta prospecting reward going deep on fewer bets and giving each enough volume to clear the warm-up. Run an S-curve like a C-curve and you’ll starve it, read the underfunded result, and kill a channel that would’ve been one of your best.

Read more at Read More

Amazon launches Alexa+ Agentic Ads

/0 Comments/in , /by

Amazon is bringing transactions directly into advertising with a new format that allows consumers to discover products, ask questions and complete purchases entirely through a conversation with Alexa+, potentially shortening the path from ad impression to conversion.

What’s happening. Amazon today introduced Alexa+ Agentic Ads, a new advertising format designed to let customers move from seeing an ad to completing a purchase without ever leaving the Alexa experience.

The format launches with partners including Papa Johns for food ordering and artists Beck, Jill Scott and Omar Courtz for concert ticket sales. The experience is currently available on Echo Show devices.

Why we care. Alexa+ Agentic Ads remove the traditional handoff between an ad and a checkout page, allowing consumers to complete purchases directly within a conversation. For early adopters, that could lead to higher conversion rates, lower drop-off and a new way to capture high-intent customers at the exact moment they’re ready to act.

How it works. Unlike traditional digital ads that redirect users to a website or app, Alexa+ Agentic Ads keep the entire purchase journey inside a conversation.

Users can engage with an ad, ask questions, compare options, check availability and complete a transaction through natural language interactions with Alexa.

The goal: eliminate friction between interest and purchase.

Concert tickets become conversational commerce. Amazon is initially showcasing the format through live event promotions.

Fans who see an ad for an upcoming concert can ask Alexa about show details, review available seats, compare pricing and purchase tickets directly through the device. Purchased tickets are then delivered to their Ticketmaster account without requiring them to open another app or website.

The experience is designed to transform entertainment advertising from an awareness channel into a direct sales channel.

Food ordering gets the same treatment. The format also extends to restaurant ordering.

A customer looking for dinner ideas could encounter a Papa Johns ad and begin placing an order immediately. Because Alexa+ can draw on previous interactions and preferences, it may suggest favorite toppings or commonly ordered meals before completing the transaction.

The entire process—from ad exposure to order confirmation—takes place within the conversation.

What to watch. Alexa+ Agentic Ads could offer an early look at how AI assistants reshape digital advertising. If consumers become comfortable completing purchases inside conversations, brands may increasingly view AI assistants not just as discovery tools but as full-fledged commerce platforms.

Read more at Read More

An open letter to everyone hiring a search leader

/0 Comments/in , /by
Search unicorn

Anthropic’s latest job posting has the SEO industry abuzz. They may as well have titled it Search Gawd. The truth is, it’s everywhere.

To be transparent, I’ve written this job description a few times and interviewed for it. I’ve yet to see any of these roles get filled, but I’ll come back to that in a minute.

Sometimes the title is Head of SEO. Sometimes it’s Director of AI Search, VP of Search, Director of SEO, AEO and GEO, or — wait for it — Agentic Commerce GEO Consultant.

Lots of titles. The assignment is basically the same: own technical SEO, understand paid search, shape content, partner with engineering and product, build measurement, prepare for AI-mediated discovery, explain it to leadership, and turn it into growth.

The predictable reaction is that this is a lot of jobs rolled into one. An entire agency behind a single employee badge. Fair, but it misses the point.

Companies have been looking for this person for years. Generative search is just forcing the issue.

This is not an Anthropic problem

This morning’s search on the job boards: 

  • Victoria’s Secret: Director, AI & Organic Search (AEO, GEO, SEO), $152K–$216K.
  • Publicis / Starcom: VP, SEO (Performance Content).
  • Accenture: Agentic Commerce GEO Consultant.
  • SailPoint: AEO/GEO Manager.
  • AirOps: Senior SEO Manager spanning SGE, Perplexity, ChatGPT, Gemini.
  • Responsive: Senior Manager, Web Strategy — SEO, GEO, plus Next.js, React, Vercel, DNS.
  • Danaher, Experian Health, Amazon News: some version of SEO + AEO + GEO.
  • Anthropic: SEO Lead, $255K–$320K.

Different industries. Different price points. Same job, unwittingly all looking for the same person.

Even the titles are arguing with the job descriptions

Agency X is hiring a “Director, SEO/SEM” whose responsibilities contain no SEO — just paid search, SEM platforms, vendor management, and a team of seven.

Consulting firm Y is hiring a “Director, SEO/AIO,” where AIO appears to be an in-house acronym no one bothered to define.

An indy agency’s “VP/Director, SEO” lists paid search, paid social, and pharmaceutical marketing among the nice-to-haves.

A token research firm is hiring a “Director, SEO & AEO” whose responsibilities actually describe SEO and AEO work — rare enough to be worth mentioning.

If the company can’t agree on what the role is before posting it, the candidate has no chance of meeting expectations that were never written down.

The taxonomy says one thing. The JD says another. The recruiter screens for a third. The hiring manager interviews for a fourth. The ATS filters out anyone worth a shit.

Looking for the missing link

You need someone who can see across technical search, content, PR, product, engineering, analytics, performance media, and brand — and understand that those functions were never as independent as the org chart suggested.

Search has always exposed the seams. A technical problem can look like a content problem. A content problem can be a product problem. A visibility problem may be an authority problem, not an optimization problem. Paid search often surfaces a messaging problem before brand research does.

Generative discovery makes those dependencies impossible to ignore. When results become answers, SEO stops being a traffic function.

At the risk of going full Yoda to avoid AI-slop speak: found, information is, only if infrastructure allows it. Content makes it understood. Brand makes it trusted. Product turns discovery into use — or it doesn’t.

You’re not asking one person to execute every task. You’re asking one person to understand how the pieces connect. That person exists. Your chances of finding that person through a conventional scoring system are slim by design.

The résumé will not look the way you expect

The value of this candidate isn’t captured by years under an SEO title or a checklist of software. The value is judgment:

  • Knowing which technical issue matters and which is noise.
  • Recognizing when the content team can’t solve the content problem.
  • Knowing when to spend, when to automate, when to wait, and when to tell leadership to stop doing that.

That judgment is hard to capture on a résumé. The candidate may have moved through agencies, publishing, product, consulting, and operating roles. Their career may look less focused than a specialist’s. That’s precisely why they can do the job.

Your ATS will screen them out. Your recruiter will flag them as “non-linear.” Your hiring panel will note they haven’t held the title before. Well, the title didn’t exist before. No one can agree on what to call it.

You can see how this search is already going sideways.

A less charitable possibility

Some of these processes may be less about filling a role than learning from the people willing to interview for it.

Senior candidates diagnose. They explain how they’d structure the function, where the organization is weak, what the first 90 days should look like, which tools they’d buy, and which work they’d kill. Invite enough of them in, and a company can collect competing organizational models and strategic priorities without hiring any of them.

Perhaps that isn’t the intent. But when a role stays open for months, gets repeatedly reposted, changes title and scope, and produces interviews that feel more like advisory sessions, candidates are entitled to ask what the company is actually buying: talent acquisition or knowledge harvesting?

The solution isn’t a shorter job description

The breadth is real, so cutting half the bullets doesn’t make the work disappear. Decide what you want. Is it:

  • A specialist who will execute?
  • A leader who will build a team?
  • An executive who can connect search, content, product, brand, and performance?
  • A consultant who can tell you which one you need?

Those are different jobs. Pretending they’re one role and waiting for a unicorn isn’t a strategy.

A closing note, since you asked

I would, however, be very good at the job. So would a handful of others who’d get screened out for the same reason.

The Anthropic job? Not getting it.

Five years under a title that didn’t exist five years ago — I don’t have them. My résumé reads like the job spec itself, in exactly the shape an ATS is built to reject. It’s an easy system to game. So easy that anyone worth their salt knows how.

The missing link is real. Generative search didn’t create it; it just made it harder to ignore. Before you hire someone to connect these systems, make sure your company can recognize them, hire them, and let them do the job.

The company that figures out how to recognize the candidate—not just write the job description—quietly wins the next decade while everyone else argues on LinkedIn about whether GEO is a word.

Read more at Read More