Adam Riemer

The pages that disappear tend to be ones written for search engines rather than readers: keyword-focused pages, programmatic SEO output, and AI-written content at scale.

• Because Search Console still shows data for those URLs, a deindexing problem can look like a ranking problem.
• His method is to sample the URLs that lost traffic, run a site: search for each, and if the page no longer appears, that is the answer.
• Then he spiders the site for shared traces, such as recurring AI phrasing in headers or programmatic templates, and sorts every page into four choices: delete it if it does not help a user, combine fragments into a fuller resource written for people, refresh a genuinely useful page with detail only someone with direct experience would know, or leave a good page alone.
• He calls this one of the easiest recoveries available, though time-consuming.

The concrete test is the site: query: take a URL that lost traffic, search Google for it, and if it is gone, you have found the cause rather than guessed at one.

The same argument shows up in his writing about content formats.

• In a Search Engine Journal column he argued comparison content, listicles, and roundups are not dead — they fail only when used to game search and AI results.
• When Search Engine Land asked him to update its title tag guide, he initially thought there was nothing new to say, then concluded a lot had changed in a year with Claude Code, ChatGPT integrations, and scaled content.

Read original

Structured data and feeds do not cover everything that makes a product page convert; the copy and the experience still shape the purchase. He also warns against companies going AI-first while treating marketing basics — PPC, SEO, email, SMS, content marketing — as optional.

He grounds the case in his own experience buying a jacket, using the purchase decision to show how the page itself, not just the structured data behind it, influenced what he bought.

Read original

LLMs customize answers for each individual user while search engines show the same results to the masses, so value comes from repeat exposure with an author or creator your buyers already trust and engage with.

• A one-off or two-off mention reads as pay-to-play and goes stale quickly.
• Repeated, relevant references signal real trust and belief in the product, and can also be reused in advertising to convert more traffic into customers.

He fielded the question directly on a podcast with hosts Jon Clark and Joe DeVita — how much should you pay an affiliate for AI search visibility — and answered that the money should buy ongoing influence with the audience you actually sell to, not a mention.

Read original

He extends the stakes beyond marketing — investors may look for company leadership through AI answers, and a job seeker whose AI record is wrong could lose a chance with a recruiter.

His peak-season observation: when competitors are out of stock and you are not, LLMs can change who gets surfaced in real time, because their outputs are not tied to a top-ten list the way traditional search results are. He says he expects to write more when there is more data and more examples.

Riemer flags the real-time LLM surfacing idea as early thinking he wants to test further, rather than a settled claim.

Read original

He shares what a prompt needs to get started and a sample prompt, but is explicit about the limits: just because a system like ChatGPT or Claude says something is bad does not make it true, misfires happen, and manual review stays part of the process.

The three scenarios he covers are parasitic affiliates, accounts more likely to generate chargebacks, returns, and fake orders, and partners that may bring in stolen credit cards or fake leads.

Read original

He says affiliate management is how his own career expanded, including the conference speaking he has done.

Read original

The first step, in his framing, is knowing what your best sellers actually are — and they may not be the items you sell in the greatest volume.

He describes what he did recently with a client and then three different ways to locate best sellers for an ecommerce store, framed around Q4, when he says every minute and dollar counts.

Read original

Riemer gives a step-by-step diagnostic instead of general reassurance.

Riemer argues that a drop during a spam update may not mean the site was devalued overall, but that specific pages — keyword-focused pages, programmatic SEO output, AI-written content — were removed from the index.

• Search Console still shows data for those URLs, which makes a deindexing problem look like a ranking problem.
• His method: sample the URLs that lost traffic, run a site: search for each, and if the page no longer appears, that is the answer.
• Then spider the site for common denominators, such as repeated AI phrasing in headers or traces left by programmatic templates, and sort the pages into four choices — delete pages that do not help a user, merge fragments into a fuller resource written for people rather than search engines and LLMs, refresh pages that are genuinely useful with detail only someone with direct experience would know, and leave good pages alone.
• Reindexing can be requested once the thin content is fixed.
• He calls it one of the easiest but most time-consuming SEO recoveries.

For brands planning a campaign

A partner program that publishes thin, scaled pages can run into the same deindexing. This post supports pitching a quality-over-volume collaboration, or a page-consolidation audit, as a joint deliverable rather than more content at higher volume.

Riemer answers plainly and shows why.

In his Ask an SEO column for Search Engine Journal, Riemer responds to a reader who asks whether updating product and service pages still matters now that AI uses product feeds and schema.

• His answer is a clear yes — page copy and user experience still shape the purchase.
• He uses his own experience buying a jacket to show how the page itself, not only structured data, influences the decision, then explains why schema, feeds, and AI are not the only things to consider.
• He also warns against companies going AI-first while dropping marketing basics such as PPC, SEO, email, SMS, and content marketing.

For brands planning a campaign

A direct anchor for co-created content aimed at online retailers. It supports a piece on product-page copy and structured data working together, with the jacket example as a concrete opening a partner could reuse.

Speaking on a podcast with Jon Clark and Joe DeVita, Riemer argues that paying for a placement or a specific article does not work.

• LLMs customize answers for each individual user while search engines show the same results to the masses, so the value comes from repeat exposure with an author or creator your buyers already trust and engage with.
• If that person mentions you only once or twice, it reads as pay-to-play and the information goes stale quickly.
• Repeated, relevant references show actual trust and belief in the product, and can also be reused in advertising to convert more traffic into customers.
• He says he expects that repeated-reference model to hold more credibility over time.

For brands planning a campaign

Shapes format and budget choices toward ongoing creator relationships with reuse rights rather than one-off placements. Useful as a stated rationale for a multi-touch partnership structure.

Riemer describes a post on three ways to use AI to help detect and prevent affiliate fraud, written in response to a reader question about low- to no-value partners.

• He shares what a prompt needs to start and a sample prompt, plus cautions: just because a system like ChatGPT or Claude says something is bad does not make it true, misfires will happen, and manual review stays in the process.
• The three scenarios are parasitic affiliates, accounts more likely to generate chargebacks, returns, and fake orders, and partners that may bring in stolen credit cards or fake leads.

For brands planning a campaign

For a software partner, a prompt-plus-human-review workflow is a ready-made co-created asset. It also shows how Riemer positions AI as assistance rather than authority, which matters when the audience is skeptical of automated screening.

Riemer frames the problem of AI systems showing incorrect or outdated information about products, services, or personal brands, and lists three things he is trying as a way to keep those outputs current: reduce support questions and cost, make sure people asking know whether the company is or is not a solution for them, and surface the company for a specific user type.

• He extends the stakes beyond business — investors may look for company leadership, and a job seeker whose AI record is wrong could lose a chance with a recruiter or hiring manager.
• He also notes that in peak seasons, when competitors are out of stock and you are not, LLM outputs can change who gets surfaced in real time, unlike traditional search results tied to a top-ten list.
• He says he expects to write more when there is more data and more examples.

For brands planning a campaign

Positions AI-answer accuracy as both a service area and a content theme. Supports a co-created piece on monitoring and correcting model outputs for retailers, with the peak-season stock example as a hook.

Riemer writes that increasing average order value, driving conversions, and moving expiring product starts with knowing what your best sellers actually are — and they may not be the items you sell the most of. He describes a recent client engagement and then three different ways to locate best sellers for an ecommerce store, framed around Q4, when he says every minute and dollar counts.

For brands planning a campaign

An example of client-work-led content that could be co-created with an analytics, CRO, or ecommerce platform, and a natural fit for retail audiences heading into peak season.

Riemer says Search Engine Land asked him to update its guide to title tags.

• He initially thought there was nothing new to say, then decided he was wrong: a lot had changed in a year, particularly with Claude Code, ChatGPT integrations, and scaled content.
• He submitted the updated guide and notes he was interested to see what the editors kept, updated, or integrated.

For brands planning a campaign

Supports thought-leadership collaborations that refresh evergreen reference content rather than produce new posts. Also shows he works with established industry editors, which is relevant when a partner wants credibility by association.

The pages that disappear tend to be ones written for search engines rather than readers: keyword-focused pages, programmatic SEO, and AI-generated content at scale.

• Search Console still shows data for them, so the drop looks like a ranking problem.
• The diagnostic is to sample the URLs that lost traffic, check each with a site: query in Google, and if it no longer appears, spider the site for shared traces such as recurring AI phrasing in headers.
• Each page then gets one of four decisions: delete it if it does not help a user, combine it with related pages into a resource written for people, refresh it with talking points only someone with direct experience would know, or leave it alone.
• Riemer calls this one of the easiest recoveries available, though time-consuming.
• In a separate column, Riemer argues comparison content, listicles, and roundups are not dead.
• They fail only when used to game search and AI results, and work when used to build traffic, sales, and brand.
• A title-tag guide update Riemer wrote for Search Engine Land was prompted by changes from Claude Code, ChatGPT integrations, and scaled content.

Sample the affected URLs, confirm deindexing with a site: query, spider the site for common production traces, then delete, combine, refresh, or preserve each page.

For brands planning a campaign

A retailer-facing software company gets a partner who can turn a common, painful problem, inherited AI-scaled pages that stop ranking, into a concrete diagnostic and cleanup framework. That is a strong hook for a co-created guide, workshop, or audit template that puts the partner's product in front of teams managing large catalogs.

Examples: Source 1 · Source 2 · Source 3

The answer is yes.

• Riemer grounds the case in a personal experience buying a jacket, then explains why schema and feeds do not cover everything that makes a product page convert.
• The column also takes aim at AI-first companies that treat marketing basics, including PPC, SEO, email, SMS, and content marketing, as optional.

Uses a first-person buying experience as evidence for why page copy and user experience still drive conversion.

For brands planning a campaign

This is the closest match to a brief about product discovery and conversion. Riemer can argue the case for the human-readable layer that sits on top of structured data, a message that complements rather than competes with ecommerce software.

Examples: Source 1

The goals are practical: reduce support questions, clarify whether a company is or is not a fit for a given shopper, and surface the company for specific user types.

• Riemer extends the same problem to investors looking for company leadership and to job seekers whose AI profile could hurt them in a hiring process.
• One observation is a peak-season opening: when competitors are out of stock, LLMs can change who they surface in real time in a way traditional search rankings cannot, because LLM outputs are not tied to the top ten links.
• On affiliate marketing in AI search, Riemer argues paying for a placement or a single article does not work.
• Brands should pay for repeat exposure, and only when the specific author or creator can influence the audience the brand sells to.
• LLMs customize answers for individual users while search engines show results for the masses, so one-off mentions read as pay-to-play and go stale, while repeated references in relevant places signal real trust and can be reused in advertising.

Riemer notes the real-time LLM surfacing theory is early and wants more data and examples before treating it as settled.

For brands planning a campaign

This gives a software partner two angles: a co-created playbook for auditing and correcting what AI systems say about a retailer's products, and a credible point of view on how to budget creator and affiliate spend when AI answers are personalized. Both speak directly to ecommerce teams trying to understand where AI search fits their marketing mix.

Examples: Source 1 · Source 2

The piece covers what to put in a starting prompt, a sample prompt, and what to be careful about.

• Three scenarios are covered: parasitic affiliates, accounts more likely to generate chargebacks, returns, and fake orders, and partners tied to stolen credit cards or fake leads.
• Riemer is explicit that an LLM calling something bad does not make it true, misfires happen, and manual review stays necessary.
• On careers, Riemer maps three paths for an affiliate manager: corporate desk jockey, specialized skill sets up to VP level, and executive leadership, plus a fourth that is not a standard career track, entrepreneurship, whether starting an affiliate company or a marketing agency.
• Riemer says affiliate management is how their own career expanded, including conference speaking.
• In a podcast interview, Riemer described the intersection of affiliate, SEO, and CRO and fielded the recurring question of how much to pay an affiliate for AI search visibility.

For brands planning a campaign

For a software company selling into ecommerce teams, Riemer can speak to program operations, fraud screening workflows, and team structure, not only content and search. That widens a collaboration beyond a single article into training, templates, or a webinar series for affiliate and partnership managers.

Examples: Source 1 · Source 2 · Source 3

The first step, in Riemer's framing, is knowing what your best sellers actually are, and they may not be the products you sell the most of. The post shares what Riemer did recently with a client and three ways to find the best sellers for an ecommerce store, timed for the Q4 season.

For brands planning a campaign

A retailer-focused software brand can use this as the basis for a data-backed co-created guide or webinar on merchandising decisions that raise average order value. It shows Riemer works from client examples and store-level data, not only search theory.

Examples: Source 1

He links to columns answering whether comparison content and roundups still work and whether product page copy matters when AI systems use feeds and schema, and notes the editors revise what he files. Compensation and contract terms are not stated.

View example

He describes initially thinking there was nothing new to say, revising the guide around Claude Code, ChatGPT integrations, and scaled content, and the publication editing and publishing it. Compensation and contract terms are not stated.

View example

Nayak stays at the measurement and instrumentation layer, writing SQL and event logic and asking other practitioners how they handle messy sites. Riemer works a level up, on content quality, search visibility, and affiliate programs.

Read original · Original 2

Aldrich's angle is platform announcements, product feeds, and paid search, seen from an agency that hires for those skills. Riemer's is organic search quality, content, and affiliate and partner programs.

Read original · Original 2

Reid aims at small business owners and offers AI training directly as a service. Riemer writes for ecommerce and affiliate teams and analyzes search and program problems rather than selling a workshop.

Read original · Original 2

The reviewed posts show work building a GA4 BigQuery attribution model that maps search terms to search sessions, orders, and revenue, and a GA4 ecommerce implementation that separates product detail pages from listing pages so the right events fire. For a software company whose buyers need to act on discovery and conversion data, that sits closer to the buyer's daily problem than a general SEO column.

The evidence covers analytics implementation and one open technical question, not editorial writing, affiliate programs, or AI answer accuracy.

• There is no reviewed example of co-created content, a published series, or a working audience, so a partnership would start from technical credibility rather than a demonstrated content pipeline.
• Adam Riemer's reviewed posts, by contrast, show repeated editorial columns and frameworks a retailer can follow.

Examples: Source 1 · Source 2

The reviewed post analyses Google's AI-driven shopping updates, including agentic checkout, selectable product comparisons, and local inventory feeds, and the other post is a hiring notice for a part-time paid search specialist working in Google Ads, Microsoft Advertising, Shopping, and data feeds. If the software company's buyers care most about Shopping performance and feed quality, this is a nearer match than affiliate or SEO recovery content.

The evidence is two posts, one of them a recruitment notice, so it shows a practitioner's view of shopping ads and agency hiring rather than a body of content.

• Nothing reviewed covers affiliate management, AI answer accuracy, or editorial publishing, and no audience or content-production record is documented.
• The tradeoff is depth: there is less reviewed work to judge a co-created series on.

Examples: Source 1 · Source 2

The reviewed posts explain Shopify's Canvas tool and ChatGPT's Try On feature in plain language and close with an offer of a one- or two-hour workshop built on a team's own emails, documents, and data, plus a direct contact address. A software company that wants a training-led co-marketing format rather than a published playbook has a documented route here.

The reviewed posts read as AI news explained for store owners plus a service pitch, and they lean on the tools rather than the retailer's own data.

• There is no evidence of affiliate marketing, SEO recovery, or technical SEO work, and the documented offer is AI training rather than co-created editorial content.
• A campaign wanting a published framework and a named editorial track record gets less of that here.

Examples: Source 1 · Source 2