Nils Berge

The battle against content bloat has taken an ironic turn: we're now building AI tools to filter out AI-generated fluff. A developer's new tool strips away the verbose padding and SEO optimisation that's drowning out actual substance across the web. Testing it on recipe sites and tech blogs reveals something fascinating - up to 80% of today's web content is dispensable filler. This points to a deeper issue in our information ecosystem. The rise of programmatic content creation has spawned an arms race between AI that generates and AI that filters. As one HN commenter noted, it's like "one guy pays AI to dig a hole, another pays AI to fill it." Beyond the technical solution, this highlights how economic incentives shape our online experience. The push for engagement metrics and SEO rankings has created an environment where authentic, concise communication becomes increasingly rare. Building better filters is valuable, but perhaps we should examine why we need them in the first place. What if we designed content platforms that rewarded signal over noise? https://lnkd.in/eS5whrwv #artificialintelligence #contentcreation #webdev #productdesign

🆕 Elon Musk's xAI has launched Grok Web Search 👀 and it's available at grok .com Only users in Australia can use it for now! Here's me 👇 using it to look for information related to December 2024 Spam Update which Google is rolling out nowadays!

13

7 Comments

There are 3 problems with pricing deliverables instead of billable hours that most agencies don't realize until after they've shifted their model. Here's what you should know... Over 53% of the largest advertisers have adopted deliverable-based pricing instead of billable hours. Source: ANA That adoption number has been increasing. Why? It's the easiest revenue model to switch to from billable time. Which is why it's also the weakest. 𝗣𝗿𝗼𝗯𝗹𝗲𝗺 𝟭: • Clients have industry benchmarks for deliverables • You price a deliverable instead of hours • Clients reverse engineer the staffing plan and tell you what your work should cost Some agencies that haven't adopted deliverable-based pricing have had the following experiences: • Client: "Do you use deliverable-based pricing?" • Agency: No • Client: "We do. And here's what we think your deliverables should cost." 𝗣𝗿𝗼𝗯𝗹𝗲𝗺 𝟮: • Deliverable-based pricing is a revenue model • Only pricing, contracts, and billing methods change • The most common agency issues remain You know the issues... Commoditized offerings: ↳ Describing the agency in the same ways as others ↳ Listing a menu of capabilities or deliverables ↳ "We work at the intersection of creativity and ..." Overservicing: ↳ Teams making concessions, diminishing margins. ↳ Inability to offer competitive salaries to hire ↳ Staff experience burnout, and creativity suffers Complexity and limited scalability: ↳ Assigning people to projects ↳ More projects = more meetings and overhead ↳ Little time for people to do actual work 𝗣𝗿𝗼𝗯𝗹𝗲𝗺 𝟯: • Custom deliverables; reinventing the wheel • High effort and time is required • Higher risk to profitability I've worked with many firms that have switched to a deliverable-based pricing model. It's a great first step away from selling time. Then, they see they should've transitioned to a productized value model where: 💥  We monetize their intellectual capital as repeatable solutions to high-value client problems their firm is uniquely qualified to solve. It's more than just a revenue model. It's a business model. • We refocus their business strategy • We shift from services or deliverables to "impact" • We refine the operating model for repeatable success In this model, you won't have the same problems: • You'll sell solutions, programs, and creative products • You can't be compared to benchmarks or competitors • Your pricing is based on your unique strengths • Teams are aligned on a repeatable approach • You simplify, forming teams around value, not projects You can still provide custom engagements (see yesterday's post), but not at your agency's expense. The revenue and profitability potential is far greater. Productizing your firm's value also creates opportunities to leverage AI to develop tech-enabled products for recurring subscription or licensing revenue. If this interests you, I'm presenting an online workshop on Sept 12. Use the link below to sign up👇

12

5 Comments

Anyone considering using Kit (formerly ConvertKit) should read this! Lessons learned from a super user: 1. I cannot preview a template before I use it. I can see the template partially but in order to see the full template and if it's the correct one I want to use, then I need to create the email with it. Then if it's not correct, I can edit the entire thing from there. Or delete it and start over choosing another template (that may or may not be what I am looking for). Even with proper naming, it is still nice to be able to see a preview before creation starts. 2. You cannot save an email as a template. If I work on an email, or even start from an existing template, and edit it. Then I want to use this same layout again, I cannot save it as a template. I can duplicate the email and swap out all the content but that puts me at risk of duplicating everything (like audience, tags, etc.) 3. You cannot name your emails. The email name is defaulted as the subject line so unless you are using 'Newsletter' or 'Nurture #1' as your subject line, it's very difficult to find the email you are looking for. 4. There is no foldering structure anywhere. I cannot organize my email campaigns by folders. I cannot organize my images by folders. So you scroll and scroll and scroll... 5. Tags tags and more tags. When subscribers come in from various sources, they all need to be tagged - if you want to see where they came from. Then you can make Rules to manage what happens when they have been tagged. 6. Landing pages have ZERO HTML capabilities. You are at the mercy of their templates. 7. Beware of 'integrations'. There are actually not that many. Majority of answers I get from their support team are 'you need a tool like Zapier'. Well....Zapier is not free so now I have an added cost plus another tool to manage. 8. Kit is advertising to your customers. If you use their landing pages with forms, the confirmation page adds a CTA for Kit. There is a large callout 'want to grow your own audience' and a button to learn more. Not only are your subscribers now getting taken away from your content, they are seeing an advertisement they did not wish to see. *everything is verified and documented with their support team as I have navigated through these challenges.

4

This is a pretty good breakdown of conversion funnel analysis. It’s a little different to the way that I work but basically the same. Good to see and explained in such detail. https://buff.ly/45vi2AE

1

Turn ChatGPT into readable content. Here’s how: ChatGPT has been clogging up feeds with sh**ty, generic content for a while now. If you’ve used ChatGPT to write, you’ve probably cringed at some of the stuff it spits out. AI can be useful, but on its own, it’s going to give you boring content that serves zero purpose. If you want posts that actually connect with your audience, they need a human touch. I wouldn’t recommend using ChatGPT like this, but I’m well aware that some people prioritize speed over quality. Here’s a quick example of how I'd take a basic ChatGPT draft and edit it into something more readable. ChatGPT Version: "In today's competitive digital landscape, utilizing AI in your content strategy can be a game-changer. AI tools like ChatGPT enable marketers to produce content at scale, generate unique ideas, and streamline the overall content creation process. With AI at your side, your content strategy becomes more efficient and impactful." Thought Process: - "In today's competitive digital landscape" A great way to lose the reader's attention in the first five words. Cut this. If you’re going to start in this way, at least make it relevant to a pain point. Bland phrases like "competitive digital landscape" don’t stop scrollers. - "Game-changer" What does this even mean? It is hype language with an empty meaning. If AI adds value, spell it out. What specific change does it bring? - "Produce content at scale, generate unique ideas": Okay, how? Be critical here. We’re talking about AI, so no, it doesn’t “generate unique ideas.” It remixes what’s already out there. Producing content at scale is the real value. Highlight that and be specific. - "Streamline the overall content creation process" If it’s not specific, it’s out. We don’t want the fluff. What does this actually mean for the reader? If it’s vague, it doesn’t add value. - "Your content strategy becomes more efficient and impactful" More fluff. Readers are looking for real benefits, not buzzwords. Edited Version: "AI can help lift the content burden off your shoulders. With tools like ChatGPT, you’re no longer staring at a blank screen trying to start your next blog post. Instead, you can quickly generate drafts, letting you focus on the tasks that are the best use of your time. ChatGPT won’t give you groundbreaking ideas, but it can cover the basics and free up time for your team to inject real value into the content. By handling the basics, AI saves time and keeps your content engine running smoothly." So, next time you let ChatGPT handle a draft, remember: it’s only step one. Take the AI’s version, break it down, and ask yourself what your audience actually wants to know. Cut the fluff, get specific, and humanise your content. That’s how you turn generic AI output into content that people will actually read. P.S. I edited this at 2am on Sunday night, so please forgive any spelling mistakes😄

5

I've seen simple FAQ tweaks double conversions. Not by answering questions, but by: - Clarifying key benefits - Speeding up decisions - Addressing core concerns Here's what most brands miss: FAQ data isn't about support. It's about understanding buying behavior. When customers ask questions, they reveal: → Hidden purchase blockers → Decision hesitations → Trust concerns The framework we use: 1. Track top questions 2. Spot patterns 3. Test solutions 4. Measure impact Result? → Better experience → Happier customers → Higher conversions Stop treating FAQs as just support. Start using them as conversion insights. Follow me for more CRO tips →

10

1 Comment

Big data = big buy in – Why user research is your strongest ally. Get stakeholder buy in for your design (easy mode): Talk their talk: Business = ROI & goals. Engineers = tech & problems. Big data wins: Show research with lots of users, not just a few. Communicate clearly: Use simple words, not fancy jargon. Lead without a title: Use your skills to convince, not boss. Find UX friends: Get others to support your good ideas. Bonus tip: Start small Get buy-in for a smaller project to show the benefits of UX design. P.S. What’s one small win you’ve achieved in UX design that helped you gain stakeholder trust? ♻ Repost to share with others. ⬆ Follow Sivaprasad Paliyath for more. 🔔 Hit the bell on my profile for more updates.

44

19 Comments

RECOVER FROM GOOGLE HCU - based on the rationale of data points from the leak. Those hit are likely to see recovery at FUTURE UPDATE points and not between updates. 1. USER BEHAVIOUR is at least ONE of the defining factors that determines if something is HELPFUL - note here HELPFUL and GOOD aren't the same thing. GOOD CONTENT isn't always helpful content! This would explain WHY good content got wiped out post SEP/MAR HCU. 2. BEHAVIOUR ANALYSIS (from click) NEEDS VISITORS - so we should do the following: 👉 Build LONG TAIL Traffic from organic, set up MICROSOFT CLARITY / custom GA4 exploration for engagement + GTM EVENTS to capture interactions 👉 Adjust your content so that you PRIORITISE what people care about above the fold - remember BEHAVIOUR indicates helpfulness, your focal query will have an intent behind it - look at what you can introduce to help users get what they came for, VIDEO, interactive media, UGC all have better capabilities at retention 👉 Use AlsoAsked to dump out PAA data for your query, export to sheets, send to chatGPT and ask for PRIORITY TRENDS based on the questions 👉 Adjust your content with an INFORMATION PRIORITY FRAMEWORK 👉 Make your content fucking hot - humans get bored easily, UX comes in here - you want UNCLUTTERED, SNAPPY, CLEAR layouts with information in digestible chunks with media/videos/tables/ugc 👉 Use Clarity + GA4 + GTM to look at IMPROVING ENGAGED SESSIONS and engagement time in seconds 👉 Track OUTBOUND EXIT BEHAVIOUR (GTM events on outbound clicks) 👉 Refine and TEST until you can sequeeze as much engagement as possible / retention 3. CONTENT 👉 Use inLinks as your content editor - optimise for END USER INTENT but ensure you have good solid topical coverage (NSR/NLP) incorporate relevant synoyms 👉 Ensure people writing content are high quality / understand content scopes and are able to articulate things using the right sentiment, language, person perspective 👉 Information prioritisation and digestibility 👉 Cite facts/statements 👉 Declutter - get rid of NON SERVED content 👉 AVOID EXCESSIVE TOPICAL DILUTION 4. LINKS 👉 Google appears to have a different storage policy for content less likely to be served - this may downvalue links from stale/stagnant content 👉 Links from fresher content where the linking source has traffic appears to be MORE likely to transfer equity (FEATURED/HARO/DIGITAL PR) 5. BRAND SIGNALS 👉 Increase BSV (brand search volume) 👉 Consistency in brand details 👉 History, about, address, establishment, team etc. 6. NAVBOOST 👉 Navigational queries - look at building a logical navigation that facilitates a relevant hierachy that's likely to result in further engagement 7. CTR Improvement 👉 Sites generating IMPRESSIONS but not clicks are likely to be demoted if they maintain a low CTR - FOCUS ON DRIVING UP CTR WHERE THERE ARE IMPRESSIONS! #seo #seo2024 #digitalmarketingtips

187

16 Comments

This is the "Missing Dimension" to integrate experimentation with the product operating model according to Marty Cagan: -> Balancing Low-Risk Optimization and High-Risk Innovation <- This is the third dimension, the other two 1. centralized vs decentralized structures, and 2. Product operating model vs Project & 'command and control' type ops. (see link to the miro board) This is my final clip of this series. Here I show Marty this pet project of mine, where I'm trying to understand patterns of how orgs build generally vs how they structure their experimentation function. (see link to the miro board) Topics Covered: - Centralized vs. Decentralized Experimentation Structures - Product Operating Model vs. Project Management Ops - Insights from Industry Leaders at Top Companies - The Missing Dimension: Balancing Low-Risk Optimization and High-Risk Innovation - Preparing for the Upcoming Conference in Austin Key Highlights: - Ben Labay's Miro board on experimentation team structures - Insights from leaders at Booking.com, Heineken, Hulu, Disney, American - Express, Spotify, and more - Discussion on the McKinsey article about the product operating model - Marty's thoughts on the importance of high-risk, innovative experiments Tags: #ProductManagement #Experimentation #Innovation #TeamStructure #MartyCagan #BenLabay #BusinessStrategy #TechLeadership

20

How does Google work in 2024? To understand SEO, you need to understand search engines. (Big shoutout to the team at ByteByteGo for their amazing infographic that inspired this post!) This knowledge isn't just for tech geeks - it's the key to unlocking online visibility for your business or personal brand. Here are the 4 simple steps that power the search engine magic: 1. Crawling ↳ Web crawlers (or "spiders") scour the internet for new content ↳ Tip: Ensure your sitemap is up-to-date to help crawlers find all your pages 2. Indexing ↳ Discovered pages are analyzed and stored in a massive database ↳ Tip: Use clear, descriptive titles and meta descriptions to help search engines understand your content 3. Ranking ↳ Results are ordered based on relevance and authority ↳ Tip: Build quality backlinks to boost your site's authority in the eyes of search engines 4. Serving ↳ The ranked results are presented to the user ↳ Tip: Optimize your page load speed to improve user experience and search rankings Did you release all the moving parts that go on behind a simple Google search? ➡️ Follow Richard Edwards for daily content ♻️ Repost to your network and share the power the search engine magic

240

121 Comments

My 2025 SEO Predictions: 1. Technical Site Setup: A fast, reliable tech foundation will be crucial as search engines prioritize efficiency to save crawling costs. 2. Context & Schema: Context-rich content and structured data will improve machine comprehension and rankings. The more context you can provide, the better machines can figure you out! 3. AI Content Drafts: AI drafts will become standard, this will become better with agents who know your brand and products, and are ready to write in your brand voice. It will be key to not forget about the information gain and provide those agents with fresh and researched information. 4. Trust Signals: Authority and trust will hinge on strong links and brand recognition. We'll be building more links than before, and apply strategic mixes of the types of links we use. 5. Links & Relevance: Links remain key, but contextual relevance will dominate. Don't search for links from irrelevant sites, this might dilute your message. 6. Generative AI Results: Expect Google to expand generative AI globally, refining user experience. 7. Personalization: SEO strategies must align with tailored search results for distinct user cohorts. 8. GPT Tracking: Tools to track AI-driven visibility in search engines like ChatGPT, Claude, and Gemini will emerge. 9. E-commerce SEO: Optimizing for Google's evolving organic shopping experience will be essential. Google US already looks almost like Amazon when searching to buy stuff. 10. Platforms Beyond Google: Success will depend on audience engagement across diverse platforms. It will be more important than ever to provide a seamless experience across the platforms that matter to your audience. 11. Big Tech in AI: Innovations from Meta, Microsoft, Apple, and Amazon will reshape search and AI interactions: Meta will integrate AI into FB, IG and WhatsApp, Apple into its devices, Amazon maybe into the shopping experience to stay ahead of Google. Microsoft will push AI through Office 365. Google through Google Workspaces, right now AI is an addon, but might be integrated into the standard package with an increase in the base pricing. 12. Input control: With all those machines creating answers, it'll be more important than ever to control what goes into them: on your site, but also other websites, authority magazines in your industry, social platforms, video content and more! 13. Voice assistants will respond like ChatGPT or better, taking voice search to a new level and widening AI adoption. 14. Reddit: Reddit will continue to grow, making it an essential part of a marketing mix to be active on that platform. It's often impossible to outrank a Reddit result in Google - the second best thing to do is participating on the ranking subreddit with your brand. Let me know your thoughts in the comments!

8

SKAN 5.0 is here. Or Is it? Still early days, but here’s what I know about Apple’s new AdAttribution Kit in simple terms. 1️⃣ At its core, it’s an evolution of Apple’s privacy-centric framework SKAN (Store Kit Ad Network) - we’re not quite calling it SKAN 5.0, yet.  2️⃣ It will support app marketers in the re-engagement/retention funnel stages. Got users who have stopped using your app within a 30-day window? You can re-engage these users with a targetted ad to get them back! Finally, the retention funnel getting some love. 3️⃣ Apparently…it will be able to support ‘cross-platform’ visibility, think being able to track postback events from other app marketplaces (maybe Google Play?). This is huge and hits the note of my prediction that omnichannel/cross-platform WILL be the new form of marketing. 4️⃣ Apps do not need to use App Tracking Transparency before calling Ad Attribution APIs and can call these APIs regardless of their tracking authorisation status. Not sure how this exactly works, does that mean ATT will become obsolete since Apple will automatically anonymise user data at the point of install anyway? Time will tell. 5️⃣ At the end of the day, Apple is evolving SKAN to help marketers better understand their data and acquisition channels. Apple did us dirty back in 2021, so are they trying to win us back? Yes. The main callout is that it will be a ‘better’ version of SKAN… so I’ll take it! Other points worth mentioning for my Mobile App Marketers out there: - As you can’t have visibility or analyse data from individual users, it’s more important than ever to meet your ad network privacy thresholds, for Meta, I believe it's around 80 App-installs/day - for other platforms, you’ll have to research. Yes, your spending will go up, but your quality of data will be higher (remember, the algorithms LIVE and BREATHE data). The more, the merrier. - There are 5 main parameters that are subject to privacy: Source ID (formerly campaign ID), Source app ID (app traffic only), Conversion Value/Coarse-grained (low, medium, high), Source domain (web only - Safari exclusive for now!), Conversion Value/Fine-grained (up to 64-values, but has higher privacy thresholds) - When is AdAttribution Kit live? It already is for some iOS 17.4 users, but is expected to have a full release with iOS 18 (Sep/Oct 2024). 📚 Keen on holistically developing yourself beyond marketing? How about creating meaningful connections while you’re at it? Finally, do you want to be a forward-thinking leader? If so, join the Futurist Marketers Club and let’s make it happen (link below)! #SKAN #ApplePrivacy #AdAttributionKit

4

3 Comments

This is hands down the most valuable eCommerce FREE resource you’ll see in 2024. Our team has spent 3 years and over £42m to create this exact structure which has proven to scale hundreds brands by more than 300% in 1 year. The template covers: → Our distinct acquisition campaign structures for scaling and testing. This ensures continual optimisation and avoids creative fatigue. → Audience segments by time retention and funnel placement. This utilises both ABO and CBO campaigns for optimal control over spend. We charge $5k/mo+ to implement this as a service. But I'm giving away all of our best knowledge for free for 48 hours. All you need to do is: 1. Give this post a like. 2. Drop a comment below saying “METASTRUCTURE”. 3. Connect with me so I can DM it to you.

1

We're going to see more and more businesses compete with Google Analytics. As cloud infrastructure continues to get cheaper and the Google Analytics brand being irredeemably hurt by GA4, more businesses will go after this massive market. Overall this is good news but its early days. Will Ahrefs become the next Google Analytics, or will Plausible Analytics, PostHog or some other player claim the crown?

7

3 Comments

Google rushed its new (in)famous product, AI Overviews, and failed badly! I saw that it was suggesting users to jump from Golden Bridge who were asking, "What to do if one feels suicidal?" Well, that would be the last thing to do! This was not it, there are other instances as well. In one instance, Google was suggesting to use Glue in the sandwich This happened because Google picks answers from Reddit; in this case, it was a bad move! All this was done to provide "RiCh CoNtEnT" to the audience Well, no matter how much we all hate keyword-stuffed or AI articles, at least they were not suggesting these bizarre things! Also, this came after the recent merger of Google with Reddit! Another thing destroyed by Capitalism! #seo #aioverviews

6

Google search is now filled with bullshit content for most search terms. This is a rough picture of the first page: - AI summary - Google Ads - Snippet - Over-stretched long-form content trying to shine through the AI hazard - AI posts One thing I have always relied on and still do: Add "reddit" to the end of search terms. With this, you get real opinions from people on what you actually searching for. Super Handy!

1

SEO in a world of GenAI answer engines. ‘Quote of the Day’ by Emilia Korczynska in a superb LinkedIn post looking at the conundrum it poses, and the potential solutions for Generative Engine Optimization (GEO). “And that's the irony: in order to be visible in generative AI tools, you need to have a lot of indexed content. But the better and hence more popular these tools get - the less clicks you're getting to your SEO content, and the less incentive you have to produce it.” https://lnkd.in/emVh-WhZ #SEO #GEO #GenAI

1

1 Comment

Customer objections are critical to overcome in order to secure a conversion. The same goes for stakeholder objections with our experimentation program. So, here’s 3 approaches I use to renew stakeholder buy-in. 1. ALIGN WITH PRIORITIZATION FRAMEWORK When contesting the validity or importance of an idea, show the prioritization matrix to stakeholders. Run them through the criteria, and show a few examples of hypotheses and how you rate them, you can even invite them to do the rating for bonus engagement. By doing so, you are using a data-driven approach to problem-solving. Separate the objections against the variation, to objections against the hypothesis. 2. TAKE OUT THE GUESSWORK As long as we discuss this idea, no one knows if it will beat the status quo. Even based on research, hypotheses are at best educated guesses until we validate them with a big enough sample. Tell stakeholders that this is an educated guess, but still a guess. Both you and them are currently guessing, and that doesn’t do any good to the website or business. We have to validate our assumptions through testing, and the statistics behind ensure that when we call this test, we will have confidence that the test results are not due to luck or any external variable. 3. REVISIT THE STRATEGY TOGETHER Maybe you need to take a step back and provide the context that is missing to stakeholders. Review the program’s 90 day KPI. Demonstrate how reaching these KPI help attain the 12 month goals. Show how each test ties into insights they want to learn about their business and customers. Connect the dots between test plan and research. Recap how the experiments and roadmap are built to solve the most critical problems that prevent customers from converting.

2