Cory Lindholm

The post gives three reasons.

• Google's AI optimizes within Google's ecosystem, tuning platform metrics rather than a client's profit margins and goals.
• Treating AI output as a black box leads to mediocre results; the value lies in interrogating recommendations and understanding why they were made.
• And if every agency uses the same tools the same way, the edge must come from elsewhere: unique data interpretation, custom automation logic and statistical methods layered on top.
• The conclusion: a shift from campaign operator to what the post calls an AI Orchestrator and Business Strategist.

The post answers its opening question — are experts and custom analytics now obsolete? — with three numbered reasons. One warning: 'blindly following AI recommendations is a path to mediocrity'; clients' specific profit margins should outrank Google's platform metrics.

The author treats platform AI as both a risk to guard against and an opportunity to stand out. A later post on agentic capabilities warns that unsupervised automation can drain spend; this one argues that new tools raise the ceiling for teams that build custom analytics and automation on top of them.

Read original

Flexible targets cut both ways: ROAS can slide as the system chases incremental orders, which is acceptable only when contribution-margin math is solid.

• Exploration happens out of sight: the algorithm tests queries the interface never surfaces.
• On a large, messy catalog, that can direct budget to low-lifetime-value products unless clean product-level signals are in place.
• Attribution still lags the new placements, widening the gap between platform-reported value and real value.
• The recommendation: let the model scout beyond the existing keyword list, then measure the gains against the brand's own margin benchmarks before scaling.

The post closes with a concrete recommendation: 'Let Google's model scout beyond your keyword list — then measure those pockets of lift against your own margin benchmarks before scaling,' followed by an offer to run a data-science sanity check before a reader flips the beta switch.

Read original

Seer Interactive data shows retail paid-ad CTR falling from 21 percent to 10 percent when the AI answer block appears.

• The post also repeats the source's caveat: AI Overviews appeared in only 7 percent of its paid dataset and 2.2 percent of impressions, so ads and AI summaries are rarely paired, but the damage is severe when they are.
• AI summaries draw from product feeds, schema and E-E-A-T snippets—areas many paid teams never touch—so a competitor's page or a hallucination can supply the answer a brand should own.
• The proposed fixes: audit feed and schema quality as you would Quality Score, track 'answer-share' alongside impression share, and cross-train on generative and answer engine optimization.

The audit names specific issues: missing GTIN and structured_title flags in Google Merchant Center, and a new metric defined in the post as queries with a citation divided by total queries checked.

Read original

The post cites McKinsey Digital's finding that executives underestimate internal generative-AI adoption by a factor of three, so guardrails may be missing without anyone knowing until after the damage.

• Complex catalogs amplify the risk: data gaps or price swings can send models off course, turning today's feed error into tomorrow's out-of-stock ad drain.
• The operational question: who guards the guardrails when spend spikes?

The post closes with an invitation to stress-test how strong a reader's guardrails really are, focusing on the gap between an AI's move and the first alert.

Read original

Cyber Monday cleared 13.3 billion dollars online last year, per Adobe Analytics, and reliable November performance depends on early creative pressure-testing, feed tuning and time for Smart Bidding to learn.

• If the first creative variant fails when CPMs spike, there is no time for a redo.
• No single split test or campaign template works across every catalog: product mix, margins and auction volatility shift faster than a flash sale.
• The edge comes from asking sharper questions sooner and acting while clicks are cheap.

The post points readers to a recorded '30-minute convo' with another practitioner, Aaron Young, rather than a generic call to action.

The same margin-and-preparation logic applies here to a calendar problem rather than a platform change. The focus shifts from bidding mechanics to creative testing and feed readiness, but the demand for evidence before spend remains.

Read original

In response to Google Marketing Live 2025, the post takes on the fear that AI announcements make experts and custom analytics obsolete, arguing the opposite for large, complex e-commerce accounts.

• It gives three reasons.
• First, Google's AI optimizations serve Google's ecosystem, prioritizing platform metrics over a client's specific profit margins and goals.
• Second, treating AI output as a black box leads to mediocre results; the value lies in interrogating recommendations and understanding why they were made.
• Third, if every agency uses the same tools the same way, the edge must come from elsewhere: unique data interpretation, custom automation logic including self-built AI agents, and statistical methods layered on top.
• Lindholm's conclusion is a shift from campaign operator to what the post calls an AI Orchestrator and Business Strategist, fluent in data science and able to translate it into profitable decisions.

For brands planning a campaign

A strong co-marketing webinar angle for e-commerce advertisers or agencies: turn AI anxiety into a message about human oversight and custom analytics, easier to promote than a product tour.

The post covers Smart Bidding Exploration, which Google billed at GML 2025 as its biggest bidding update in a decade.

• It lets AI flex Target ROAS in real time to capture conversions static targets miss.
• Lindholm focuses on the trade-off, not the announcement.
• Flexible targets can push ROAS downward as the model chases incremental orders—acceptable if contribution-margin math is airtight, risky if that number is a guess.
• Exploration also happens beyond what the interface shows: the algorithm tests queries the UI never surfaces, potentially directing budget toward low-lifetime-value SKUs in a large, messy catalog unless it receives clean product-level signals.
• Attribution still lags the AI, widening the gap between platform-reported and real value.
• The recommendation: let the model scout beyond the existing keyword list, then measure those pockets of lift against the brand's own margin benchmarks before scaling.

For brands planning a campaign

A concrete webinar or newsletter angle for performance teams: a sanity check before flipping the beta switch, not another general talk about AI changing PPC.

This post brings the measurement problem together.

• Google rolled AI Overviews out to every U.S. searcher, and overall click-through rates fell roughly 30 percent, according to BrightEdge.
• Seer Interactive data cited in the post shows retail paid-ad CTR dropping from 21 percent to 10 percent when the AI answer block appears.
• The same source says AI Overviews appeared in only 7 percent of its paid dataset and 2.2 percent of impressions: ads and AI answers rarely appear together, but when they do, paid CTR nearly halves.
• Separately, 58 percent of searches were already zero-click, per Search Engine Land.
• AI summaries draw on product feeds, schema, and E-E-A-T signals—areas many paid teams never touch—so a competitor's page or a hallucination can supply the answer a brand should own.
• The post proposes three fixes: audit feed and schema quality as you would Quality Score, starting with missing GTIN and structured_title flags in Merchant Center; track 'answer-share'—queries with a citation divided by total queries checked—alongside impression share; and cross-train on generative and answer engine optimization.
• It closes by asking which visibility metric matters most in a zero-click world.

For brands planning a campaign

The answer-share metric gives a webinar a named, teachable framework participants can apply the same week—more promotable than a trend summary.

The post opens with the line 'AI to delegate, not disconnect' and explains why unsupervised automation is dangerous at catalog scale.

• Agentic capabilities were the headline launch at GML 2025, but Lindholm's point is a mismatch in speed: AI makes decisions at SKU speed while a profit-and-loss statement reacts at human speed.
• A McKinsey Digital reference supports the argument that C-suites underestimate internal AI adoption by a factor of three, leaving guardrails potentially absent until damage appears.
• Complex catalogs amplify the risk: data gaps or price swings can send models off course, turning today's feed error into tomorrow's out-of-stock ad drain.
• The post asks who is guarding the guardrails when spend spikes and offers to stress-test readers' guardrails by direct message.

For brands planning a campaign

A fit for e-commerce advertisers and agency leads looking for a risk-and-oversight angle rather than another AI-optimism pitch; the stress-test offer makes a natural call to action.

The post makes the holiday case with numbers and a timeline.

• Cyber Monday cleared 13.3 billion dollars online last year, per Adobe Analytics, yet most brands still treat Q4 like a light switch flipped the week before Thanksgiving.
• Pressure-testing creatives, tuning feeds, and giving Smart Bidding time to learn before November can build momentum instead of relying on guesswork.
• If the first creative variant fails when CPMs spike, there is no time for a redo.
• The post cautions that no single split test or campaign template works across every catalog: product mix, margins, and auction volatility shift faster than a flash sale.
• The edge comes from asking sharper questions earlier and acting while clicks are still cheap.
• It credits a '30-minute convo' with Aaron Young and points readers to a link in a repost.

For brands planning a campaign

A ready-made seasonal calendar hook for a co-marketing partner: a pre-Black-Friday readiness session scheduled well before November, framed as preparation rather than prediction.

The post argues that paid teams should audit product-feed and schema quality as they do Quality Score, starting with missing GTIN and structured_title flags in Google Merchant Center; track "answer-share" alongside impression share; and cross-train in generative and answer engine optimization. It also quotes Seer's caveat that AI Overviews appeared in only 7% of the paid dataset and 2.2% of impressions, concentrating the CTR damage where ads and AI summaries appear together.

Uses named third-party studies with reference links to build a measurement case, then turns it into a short audit list: feed/schema check, answer-share metric and cross-training. Closes by asking which visibility metric readers will rely on.

Distinguishes industries with long, complex consideration cycles, such as 90-day B2B sales, where high-intent clicks still carry weight.

For brands planning a campaign

Directly usable for a webinar on AI-era PPC strategy: gives specific measured claims—CTR declines, zero-click share and AI Overview frequency—and pairs each problem with an action brands can take.

Examples: Source 1

The first gives three reasons: Google's AI optimizes within Google's ecosystem rather than for a client's profit margins; blindly following AI recommendations produces mediocre results; and the remaining edge comes from custom data interpretation and automation logic.

• The second warns that AI now makes decisions at SKU speed while a P&L reacts at human speed, cites McKinsey's finding that executives underestimate internal generative-AI adoption by a factor of three, and argues that complex catalogs amplify risk because feed gaps or price swings can send models into unprofitable territory.
• The stated remedy is to name a guardrail owner who can be held accountable when spend spikes.

Uses numbered arguments and analogies—a toddler given espresso and car keys—to make the oversight case readable. Each post ends with a direct offer to stress-test a reader's guardrails over direct message.

Frames AI as an opportunity to differentiate by building custom analytics and automation on top of platform tools, not just as a risk.

For brands planning a campaign

Supports a co-marketing webinar on operating AI-driven campaigns safely in large e-commerce accounts, and shows the creator will engage one-to-one on guardrail reviews.

Examples: Source 1 · Source 2

Flexible targets can let ROAS slide as the system chases incremental orders, which is safe only when contribution-margin math is solid.

• Exploration tests queries the interface never surfaces, so a large, messy catalog can burn budget on low-lifetime-value SKUs without clean product-level signals.
• Attribution lags behind the new placements: ads now appear inside AI Overviews and AI Mode, while post-click data has not caught up.
• The recommendation is to let the model scout beyond the keyword list, then measure gains against the brand's own margin benchmarks before scaling.

Weighs one platform feature by naming its upside and three concrete failure modes, each tied to a business metric. Invites readers with complex feeds to request a data-science sanity check.

For brands planning a campaign

Well suited to a technical audience evaluating AI bidding changes: takes an accountable position rather than recapping features, and pairs the analysis with an offer to inspect another team's data.

Examples: Source 1

It warns that no single split test or campaign template fits every catalog because product mix, margins and auction volatility move faster than a flash sale. The post points readers to a 30-minute conversation with another practitioner, Aaron Young.

Combines seasonal urgency with a clear constraint—no universal template—and points readers to a recorded conversation rather than a generic call to action.

For brands planning a campaign

Useful for agencies and e-commerce teams planning Q4 spend; also shows a willingness to co-present with other practitioners, a fit for a partner-led webinar.

Examples: Source 1

They disclose no sponsorship, employment, or client relationship with Google. Marketers considering a partnership should ask about any ties.

View example

Josh Gregory's posts cover scaling problems across Google and Meta, channel overlap, funnel gaps, creative testing, and margin checks.

Josh Gregory stays close to account diagnostics and audit offers for DTC brands; Cory Lindholm focuses more on platform AI changes, bidding logic, zero-click measurement, and feed or schema quality.

Read original · Original 2

Muhammad Salman Arshad writes from the founder and operator side—logistics, repeat purchase, burnout, and the move from freelancing to an agency—while Cory Lindholm focuses on paid search, Google Ads automation, and AI-era measurement.

Read original · Original 2

Rupert Morris addresses B2B marketing and AI tool adoption broadly; Cory Lindholm focuses on paid search execution, Google Ads bidding, feeds, and guardrails for e-commerce accounts.

Read original · Original 2

The available posts identify failures that emerge once Meta or Google budgets pass roughly £25–30K a month: cold-audience funnel gaps, Google and Meta planned in isolation, and no lifetime-value or offer strategy.

• They also give structural fixes: weekly creative testing cycles, margin and customer-acquisition-cost checks before budgets increase, offers that work for first-time buyers, funnel mapping from awareness to purchase, and landing pages built for cold clicks.
• The second post addresses a symptom many audience members will recognise—spending more without sales moving—and attributes it to targeting widening and conversion rates falling when spend rises too fast.
• It offers three remedies: campaign structure that controls where budget goes, audience signals and exclusions to keep targeting sharp, and search terms, feeds, and creative aligned to intent.
• The post closes with an offer of free Google Ads audits to e-commerce brands, giving a co-marketing partner a natural joint offer to promote.
• For a software product focused on account structure, creative testing, or cross-channel planning, this is a closer fit than a session on bidding features.

This is operational advice, not AI-era strategy.

• Neither post discusses AI Overviews, answer-share, agentic bidding, or Smart Bidding Exploration, so the host would need to supply the Google Marketing Live framing and the argument about measurement under AI search.
• The claims are broad and unsourced: no named studies, account examples, or figures showing how many brands face these failures or what results followed a fix.
• Most of the detail is a bulleted pattern list rather than a worked case, and the audit offer gives no format, fee, or past-audit outcome.
• A software company seeking a guest with third-party research and a measurable framework would find structure and diagnosis here instead.

Examples: Source 1 · Source 2

The first post diagnoses why e-commerce startups fail: margins that were never modelled, illustrated with a Rs.

• 3,000 sale against Rs.
• 1,500 acquisition and Rs.
• 500 operations and delivery; logistics and third-party delivery failures that damage the customer experience; creative fatigue from running one good ad into the ground; founder burnout from one person covering product, marketing, logistics, service, content, and hiring; cash-flow timing against gateways, returns, and supplier payments; and no repeat customers.
• It ends with three self-check questions on profitability after returns and shipping, control over fulfilment, and surviving 60 days without sales—ready-made workshop material.
• The second post covers the shift from Fiverr seller to agency owner and credits LinkedIn visibility, paid and organic content, and higher-value long-term clients with the change.
• It suits a session for agencies or consultants on client acquisition and positioning.

The material steps away from paid search and AI almost entirely.

• Neither post covers Google Ads, feeds, attribution, or answer-share, so the host would need to supply the AI-era PPC frame.
• The startup-failure post describes patterns without a named source or real account, and the rupee figures are illustrative rather than measured.
• The agency post is motivational, listing before-and-after outcomes without supporting numbers; it points readers to a named contact for a roadmap, leaving the practical detail outside the post.
• Two short posts offer little basis for judging how Muhammad Salman Arshad would handle a 45-minute joint session.
• Ask for audience size, engagement rate, and past-webinar results before booking.

Examples: Source 1 · Source 2

The first post is a practical AI-adoption briefing.

• It describes changes brought by two new model releases, including working inside email, docs, and CRM without copy-paste; following house templates so decks and documents come out in the right format; and Claude becoming the default model inside Slack and arriving in Salesforce.
• It then recommends a concrete first step: take the team's most repetitive task, use the real template and real data, see how close the model gets, and reinvest the recovered hours in client and team work.
• The second post tackles zero-click search for B2B marketers: over 68% of searches in the first four months of the year ending without a click; buyers discovering, comparing, and forming opinions of a brand without visiting its site; attribution showing only a late interaction in a much longer journey; and the case for measuring something other than a linear path.
• This is the same question Cory Lindholm's AI Overviews post raises, applied to a services audience rather than a retail one, so the two could share a panel on measurement without repeating each other.

This material is furthest from paid-search operations.

• Neither post covers feed quality, Merchant Center flags, Target ROAS, contribution margin, or campaign structure, so the host would need to supply the argument about guardrails inside the ad account.
• The posts offer little detail: the AI briefing lists capabilities without tests, screenshots, or measured time savings, and says nothing about where the tools fall short.
• The zero-click post promotes a downloadable analysis but does not include the research method, underlying data, or specific recommendations; buyers cannot verify the 68% figure or see what measurement replaces attribution.
• Cory Lindholm's AI Overviews post names BrightEdge, Seer Interactive, Search Engine Land, McKinsey, and Google, while these posts refer to their own research without publishing it.
• Ask for evidence supporting the 68% figure and access to the underlying analysis before booking.

Examples: Source 1 · Source 2