Kyle Hutton

Something SGE related I haven’t seen anyone talk about: the individual components that make up these results are all tagged within data attributes. Take a look yourself by loading up any SGE result that returns a paragraph, inspecting the element and searching for "SGEParagraphFeedback". I think this is cool for a few reasons: 1. We gain an understanding of what Google considers to be a distinct feature.  2. Scraping using these attributes should be subject to less change than random divs and spans. 3. We can use them to detect the presence of specific components within the results. 4. We can use their presence – or the presence of different combinations – to classify feature intent and how that contrasts with the vanilla results. You can find the 17 I’ve found listed below 👇 #seo #ai

213

19 Comments

Struggling with inconsistently filled out experiment test plans? Liftmap just added a new feature called "Data richness" which offers a gentle reminder (and light gamification 😀 ) to teams to add all of the actual data/fields we deem important at Conversion for the completion of a well-documented test plan. A few highlights and why we think they are important Description information (Problem Statement, Objective) - These questions are focused on rich text data describing what it is that the experiment is actually trying to answer. Experiment-level Data (Size, Risk, Channel, Zone/Page, Device, Prioritization Score) - These are used for running successful queries to find completed experiments later and to uncover interesting and surprising patterns. (Ex. Query all homepage tests) Variation-level Data (Images, Hypotheses, Levers, Evidence Source, Design components) - These are the most important items collected in my opinion and highlight what changes are actually being made. This data is what allows someone a rich understanding of the test at a later date (accurate screenshots) and also what allows successful pattern level meta-analysis to be conducted (Ex. Query all CTA tests, all Social Proof tests, etc). Results Data (Variation Outcomes, Insights, Overall result, Next steps, Key finding) - These fields are a combination of rich-text and simple winner/loser/inconclusive fields for queries. Used to provide context to an outcome. Of course we tag much more than this... But these are the non-negotiable minimums for us. Curious what others are tagging to create a robust and organized system that can be referenced by broader product teams and/or used for meta-analysis? Anything surprising in here? Jakub Linowski Sim Lenz Luis Trindade ☕️Aaron Gasperi Lukas Vermeer How do you handle this at your organizations?

28

5 Comments

Loved talking to Mark Tungate for AdForum about data as an art as well as a science: how data at its best is as much about the linguistic, emotional and cultural as the numerical. We discuss how Data with Soul at AMV BBDO is ultimately about data with EQ, as well as IQ. When it came to choosing what to study, I didn't know that Computer Science, Data Analytics, Machine Learning or any adjacent degrees were even an option. I studied French Language and Literature, and I rely on the skill set I developed through this choice as much as I do on the technical skills I learnt later: the ability to analyse language, from word choice and tone to the most recurring themes and feelings in a body of text, and how they all intersect. (Although, I do wish I'd known how to code at that point, it would have sped my essays up quite a bit.) No matter the source or data set - whether it’s social listening, surveys, searches, or 19th century novels, or the output - a report, an essay, an article - they all require the same ability to analyse meticulously, construct the most important narrative and make our insights clear and actionable, whilst positioning them in cultural context. Studying literature and language arms us with the ability to unpick what a source teaches us about the socio-political climate of the time in which a piece of work was written or created- and the meaning the author wanted to convey. I've been reminded of this more than ever in the last couple of weeks.   I started my career not only alongside the wake of Trumpism and Brexit, but because of it. I wanted to understand how data had been inaccurately gathered or unfairly represented: the missing sources, the asterisks ignored, the time frames and sample sizes in tiny fonts, and the context left out. How the siloing of meaning into unrepresentative numbers saw statistics turn to slogans that fed misinformation and disinformation. When the distinctly human and contextual elements of gathering, analysing, presenting and understanding data are overlooked, the meaning and messages conveyed are at best unimaginative and reductive, and at worst, powerfully misleading. Both in perception and in practice, both inside and outside of the industry, we should remember that data can direct emotion and culture as much as it brings them to light. Above all, data is most impactful as storytelling, designed to make its readers think, feel and act.

69

2 Comments

When communicating data, context is 👑 check out my pixel helping me put tomorrows weather into context by telling me how much colder it will be compared with today. 👏 #datacommunication #datastorytelling #dataviz #dataanalysis

21

4 Comments

In the past few weeks I have been contacted by a number of university graduates asking for some guidance on the best way to get a role in the loyalty industry. While my perspective is only one, the questions I was asked made me reflect on some defining moments in my last 31 squiggly years of being in the direct marketing and loyalty industry. These have helped in some way to keep me doing what I love, and loving what I do … Here are five ideas I shared with these aspiring loyalty professionals: 1.Join every possible loyalty program you can. About 13 years ago, for 10 days in a row, I walked the streets and shopping centres asking every retail store if I could join their loyalty program. Depending on what they said, I would have a few followup questions (if no, why not, if yes, what do I get from being a member). I captured every interaction into a spreadsheet and soon compiled some learnings that gave me a foundation of what I thought was a bad, good, great experience. I also went online and added non-retail based categories. I do repeat this exercise periodically as the speed of change is extraordinary and the industry is growing rapidly. 2. Create loyalty content (whichever medium you like)  I write everyday something about loyalty and loyalty programs. It’s my loyalty journal. Sometimes a few words. Sometimes a few thousand words. The purpose is to help with embedding ideas into my memory and if I share these as articles publicly, then I am also aiming to help others. 3. Consume loyalty content There is an explosion of content from around the world - books, podcasts, blogs, vlogcasts, webinars, newsletters, interviews, research, events, conferences, meetups, articles and educations courses. There is an abundance. Find those that inspire. 4. Go deeper on the loyalty landscape, there are niches within niches and find an area of fascination Psychology, behavioural economics, gaming, finance, technology, data (woweee so much there)…and more 5. Brands you love who have a loyalty program Create a list of brands across all categories which you would love to work with and join their program and be an active member. If you happen to get an opportunity there, present your view on their program… the good, the bad, the brilliant! Finally be clear on your purpose for focusing on the loyalty industry. Why does it appeal to you? (I think as you go through the above five, it might help to add clarity...) The reasons I love the industry (from the consulting side) is the opportunity to have so much variety in the brands I am lucky to partner with, the problems I am working to solve and the science+psychology of the industry...and one of the articles I wrote about 4 years ago reflects on that... #loyalty #loyaltyprogram  #loyaltyprograms https://lnkd.in/gaJR-ji

67

16 Comments

Weekends are for progress. Dashboard updates incoming 🤝🏼. Data is key. Interpreting that data correctly is the lock. I can’t get enough of visualising data. I find it increases my efficiency massively if done right. One of my favourites is color indication for priority. Conditional formatting makes my life so much better and I can identify key areas quicker to extrapolate 👍🏼. What are some of you’re key metrics that you love visualising?

4

If SGE and how to understand and be prepared for it is gnawing at your brain (as it does mine) this is a great resource that's definitely worth bookmarking! #sge

1

What are the alternatives to Google Analytics 4? We know there is Adobe Analytics, yet increasing Amplitude Analytics is making waves in the industry. It is important to note, that both have different purposes; GA4 is geared towards web and app analytics, whilst Amplitude is focused on product analytics. That said, let's still have a look at GA4 vs Amplitude: ⚫ Features 📗 GA4 - analyse user acquisition, traffic, trends 📘 Amplitude - analyse product usage, journey impacts, adapt products to users ⚫ User Interface 📗 GA4 - complex interface that is very different to Universal Analytics 📘 Amplitude - clear user interface, better filter options ⚫ Integrations 📗 GA4 - walled garden approach, offering rolodex of Google suite products to plug in 📘 Amplitude - open approach, over 100 integrations available including Snowflake, Amazon Web Services, Google Big Query and more ⚫ GDPR & HIPAA Compliance 📗 GA4 - not compliant in a number of countries due to data transfer to the U.S and not HIPAA compliant 📘 Amplitude - announced it is GDPR and HIPAA compliant ⚫ Data Collection 📗 GA4 - Event based and user-level data collection 📘 Amplitude - Product analytics focus ⚫ Pricing 📗 GA4 - Free and 360 available, latter c.$150,000 per annum 📘 Amplitude - Starter (free), Growth and Enterprise available. No pricing available on-site Do you use Amplitude? What do you think? #ga4 #amplitudeanalytics #analyticssoftware #userbehaviour #digitalanalyticsconsultant

7

1 Comment

I've been asked several times recently about what a typical week looks like for me, and I've noticed similar posts here. So, here's a peek into some highlights of my upcoming week at Ruche Marketing LLP with Rachel Craig: ➡ Monday: Reviewing dashboards and analytics to assess last week's performance and ensure everything is set for the week ahead for our clients and business. We’re also conducting interviews. ➡ Tuesday: A 1-2-1 session with my business coach, Marie Cowgill, at the NatWest Accelerator Programme, to help develop and grow our business. ➡ Wednesday: Focusing on business development and exploring new ideas. ➡ Thursday: Presenting the Marketing Essentials Workshop in Ashton, part of the Build A Business programme from Manchester Libraries. ➡ Friday: A day for personal development, admin, and catching up, plus preparing for a big event next week (more details to come). This blend of strategy, learning, and client focus keeps each week exciting and productive. What does your week look like? #WeeklyHighlights #EntrepreneurLife #Productivity #WorkLife

14

1 Comment

'Insights in search of a decision' OR 'Decisions in search of Insights'? It is almost the chicken or the egg story but a recent post on this platform has got me thinking. Most data teams are setup (as a start) as a reactive element. A decision needs to be made and leadership is in need of insights to make the best decision possible. When the team matures both the team and the leadership calls out the need/wish to move to a proactive model. "Tell us what we dont know". This is all to familiar for most of us, but here comes my question: When you switched to proactive insights did you also change the type of decisions you were influencing? My hunch is that most of us will have to relucantly say "No" to this question. Even when data teams become proactive they often still pre select their insights on the types of decisions their leadership wants to make but not on the decisions the data tells them to make. Any that can eithet confirm or deny my hunch? #dataanalytics #dataorg #Insights

6

DR and DA are not ranking factors 🛑 A couple of people I've spoken to this week have bought this up, so I thought a post about this could be useful 👩🏽‍💻 DR and DA are metrics made by #Ahrefs and Moz to determine the "trustworthiness" of a domain (in a roundabout way). ❌ There’s no evidence that search engines use Domain Rating (Domain Authority, the power of the domain, etc) as a ranking factor at all ❌ That doesn't mean these metrics are completely useless. They can help to create a picture of a website's "trustworthiness" and are a good (ish) predictor of how well a website will perform in #SERPs. They also offer a way to compare websites within the same niche or industry. But don't believe they directly impact how well a website performs on the SERPs. They also shouldn't be used to determine the "quality" of a #backlink. But that's a post for another day 🫣

44

5 Comments

Here's what I've come to realise... The more data you have, the less you realise you know. Sound crazy? It's the truth behind data maturity. I've seen it time and again with clients: 1️⃣ Platform Intelligence You think you have so much data. Every bloody channel has its own metrics. You feel like a proper data wizard. Hate to break it to you, but you're stuck in a silo. 2️⃣ Omnichannel Intelligence Nice one, you've connected the dots. Customer journey all mapped out. Pat yourself on the back. But hang on... where's the commercial impact? 3️⃣ Commercial Intelligence This is where it gets interesting. You've linked marketing to money. Your CFO might even buy you a pint. And then it hits you: bloody hell, I knew nothing before. It's like climbing a mountain, only to realise you were just at the foothill. The higher you go, the more you see what you missed: Platform Intelligence: "We know it all!" Omnichannel Intelligence: "We knew nothing before!" Commercial Intelligence: "We're just beginners, aren't we?" 👉 The point? True marketing wisdom is admitting how much you don't know. I learned this the hard way. Thought I was hot stuff until I linked everything (including brand) to the bottom line. So, here's my challenge to you: admit you don't know it all. It's scary, but it's the only way to get better.

6

What if I told you… I wouldn’t do it again? YouTube has been incredible for building my authority in digital analytics—175k subscribers and countless opportunities to help people solve their analytics challenges. But would I do it all over again? No. I loved it—but it was brutal. Creating long-form content demands dedication and consistency. And honestly, it’s only sustainable if you can make it feel effortless. Today, I believe there are smarter, faster ways to build credibility. One of them? Short-form video. You might be thinking, “But isn’t that still a ton of work to produce?” Not exactly… The trick is to make it effortless. One way that works great for me is when I go on podcasts as a guest, I ask the host for the footage. Then, I repurpose that into short clips for YouTube, LinkedIn, Instagram, and more. All from just one interview. But what if you’re not getting invited to podcasts? That’s what inspired a project I’m excited to share. 🎥 We are developing an app that interviews you—helping you produce short-form video content for your social media just by answering questions. And now we are looking for beta testers. Interested? Comment INTERVIEW ME and be one of the first to try it out! P.S. Have you tried to create short-form video?

109

44 Comments

Today has highlighted a very good reason (again) why my clients like having me in their corner. My clients typically aren't data experts, and they need someone to understand the data they are buying. Particularly in cases where they don't appear to be getting what they thought they were buying. The project I am working on at the moment is slightly frustrating as I had done a full ranging analysis based on the data I was able to access through the data provider's tool. However, as more information became available, I realised that a few key products hadn't appeared on the ranging list, so I decided to query it. The data provider is insisting that my list (a few hundred products) contains all the products they have on their list (over 1,000). While their list contains all the products I can see, I cannot see all the products on their list. All of these back and forth emails, the excel know how to speed up a cross check, it all takes expertise and time that my client won't always have. My word of advice - don't take the reports you receive at face value, if something feels off, then check. And call in an expert if you think you need some help.

10

4 Comments

A fancy iceberg viz and a 3-step model! Don't say I don't treat you. I'm gonna sound like a broken record ↓ The way we work and how we think about working with digital tech needs to adapt to capture value from Gen AI. It is not a 'understanding how to use the tool' alone solution. This a new framework for a new time. Just like we changed when the internet hit, and when social media hit us later, we changed how we worked, lived and engaged with each other. This image from another amazing Boston Consulting Group (BCG) report (you're killing it folks) highlights this change management piece. For L&D folks, we're one part of this puzzle. But all of this only really works if all parts of your organisation are sailing in the right direction together to maximise the full potential of Gen AI. #learninganddevelopment #education #artificialintelligence

23

9 Comments

Sundays are good days to review your data. ESPs deliverability dashboards are curated based their metrics. Their ratings you see aren’t the full picture. Every week run a deliverability test: ⚒️ Check overall domain health -MXtoolbox has a free checker ⚒️ Check blocklists - you can use multirbl (.) valli (.) org ⚒️Check your DNS setup and diagnose tech problems with aboutmy (.) email ⚒️ Review your optins and automations metrics (low opens, high bounces can indicate low quality traffic) ⚒️ Review your list segmentation(s) and make sure you have a good cleaning process in place. At the end of the day, your ESPs deliverability is only one feedback loop. Don’t follow it blindly.

13

1 Comment

You know those little popups when you land on a site that says sign-up for 10% off your first order? You're about to see a whole lot more of them moving forward as websites pivot to prioritising first-party data collection. Easier for e-commerce websites, but lead-gen are going to have to get particularly inventive with how they can entice users to part with their info. #GA4 #GoogleAds #Remarketing #Analytics https://lnkd.in/eVmyvjW6

15

How is this Junk data industry sleeping at night. Not all are bad but plenty are just peddling data they know is rubbish! Thanks for sharing Arielle and thanks Suzanne Smalley for writing this important piece.

2

A big value add to my data science skill set was thinking in terms of "the system" rather than "the model". Focusing more on where the inputs come from and the outputs go is the difference between GPT3 (of interest only to other data scientists) and ChatGPT (11th most popular website in the world). From a tech POV this means: * Understanding the basics of Cloud Engineering * Data scraping and database skills * Scheduling! Even as something as simple as cron can really upgrade the value of a script. * Turning scripts into APIs and Command Line utilities From a engineering planning POV this means: * Understanding Dependencies * Designing for Modularity From a relationships POV this means: * Developing stronger relationships with whoever originates the data you're analysing * Developing stronger relationships with whoever consumes your analysis From a modelling POV this means: * Asking what new datasets could be ingested to predict this problem. * Reframing and negotiating the problem statement (e.g. are we doing classification or regression?)

9

2 Comments