Jep Castelein’s Post

Recently, a Marketing Operations person went on parental leave and gave their handover document to AI instead of to a human. You won’t believe what happened next... Sorry to disappoint you, but nothing happened because AI MOPS doesn’t exist yet. But it does give a glimpse of the future, and the future isn’t too far away. When I wrote my post about the 3rd era of Marketing Automation being AI-powered, Steven Moody had some great thoughts in a repost (https://lnkd.in/gg9kwjud). Here is the first point he made:. “𝘠𝘰𝘶 𝘥𝘦𝘴𝘤𝘳𝘪𝘣𝘦 𝘺𝘰𝘶𝘳 𝘣𝘶𝘴𝘪𝘯𝘦𝘴𝘴 𝘵𝘰 𝘵𝘩𝘦 𝘈𝘐. 𝘍𝘰𝘳 𝘮𝘢𝘳𝘬𝘦𝘵𝘪𝘯𝘨 𝘢𝘶𝘵𝘰𝘮𝘢𝘵𝘪𝘰𝘯, 𝘺𝘰𝘶 𝘴𝘩𝘢𝘳𝘦 𝘺𝘰𝘶𝘳 𝘢𝘤𝘵𝘶𝘢𝘭 𝘣𝘦𝘴𝘵 𝘤𝘶𝘴𝘵𝘰𝘮𝘦𝘳𝘴, 𝘺𝘰𝘶𝘳 𝘤𝘶𝘳𝘳𝘦𝘯𝘵 𝘸𝘰𝘳𝘬𝘧𝘭𝘰𝘸𝘴 𝘢𝘤𝘳𝘰𝘴𝘴 𝘵𝘰𝘰𝘭𝘴, 𝘺𝘰𝘶𝘳 𝘵𝘦𝘢𝘮 𝘴𝘵𝘳𝘶𝘤𝘵𝘶𝘳𝘦, 𝘺𝘰𝘶𝘳 𝘪𝘯𝘥𝘶𝘴𝘵𝘳𝘺 𝘤𝘰𝘯𝘵𝘦𝘹𝘵, 𝘢𝘯𝘥 𝘺𝘰𝘶𝘳 𝘨𝘰𝘢𝘭𝘴.” - Steven Moody ChatGPT can remember little chunks of information if you ask it to. Now imagine a 3rd-generation marketing automation system ingesting a comprehensive briefing on your company, the product(s), the market, key differentiators, the marketing and sales teams, the marketing goals, etc. It can also read your website, your marketing materials, reviews on G2, you name it. Think about it as training a new team member. Using that wealth of information, an AI MAP can start making recommendations on which campaigns to develop and what goals to strive for. Steven also mentions that the AI MAP could learn from aggregated data from many different customers. That would be awesome, but I do think it will be tricky to get approval from companies to essentially train a generic AI model using their data, and then share it with all other system users. Clearly, lots of things still need to be figured out. Where do you see the key opportunities and the key problems with AI MOPS?

The end is near for the 2nd era of Marketing Automation Wait, if we’re in the 2nd era, what was the 1st era? I stumbled onto one of my old blog posts from 2010 with a list of then current Marketing Automation systems: https://lnkd.in/gW4VE9cG. That was a transformative period, where the 1st generation vendors were still around, but the 2nd generation vendors were up-and-coming. 1st era (1999-2010): Eloqua, Marketo2Lead and Vtrenz 2nd era (2010-2025): HubSpot, Marketo, Pardot and Act-On 3rd era (2025-): who knows, but it’ll be AI-powered! Most of the 2nd gen MAPs are stalled. They are no longer innovating, their customer service is at most adequate (usually worse), and they’re raising prices just because they can. The only exception is HubSpot, which has steadily been building out their product suite, and limiting price increases. At the same time, AI will - at some point - revolutionize all software, including Marketing Automation software. I’m not talking about an automatically generated subject line, that’s just eye candy. I’m talking about a fundamental rethinking of how campaigns are ideated, planned, executed, evaluated and optimized. I can’t list the vendors yet, because no full-featured next-gen MAPs have launched yet. But I have the feeling that - in garages all over the world - people are working on them. It’s just a matter of time until they come out of the woodwork. What do you think, are we at the verge of a new era, and what do you hope the new vendors will look like?

Freshness of data! I just got a call from a tech vendor asking about my role at Backbase, a company I left 15 years ago. I don't really know what to say about that. They could have done a LinkedIn lookup. They could have asked a research team to validate old contact information. They could have used Clay to automatically validate their list. The more I think about it, CRM and Marketing Automation systems should have built-in features to improve data quality. It's not surprising that Apollo is built on top of data, and HubSpot acquired Clearbit. I'm curious what a new generation of tools will do to improve data quality. Anyone have a wish list?

I just learned that several of my favorite marketing people are speaking at Martech World Forum in Santa Clara on Aug 7-9. Sounds like a cool conference to go to. Here are my faves: Jon Miller, Scott Brinker, Darrell Alfonso, Mike Rizzo, Juan Mendoza And these are the themes: • The Future of Customer Data • The Future of the Martech Stack • The Future of Content and Experience • The Future of Operations More details on https://lnkd.in/g_6H_87Y PS. I hope people realize this is not actually in San Francisco 🙂

As I start a new chapter, I want to thank all my incredible clients and partners who have supported my work as an independent marketing operations consultant for the past 5 years. Your trust and collaboration have been much appreciated and I'm very proud of what we've accomplished together! Special thanks to the early clients who helped get my business off the ground: Alicia B., Alexandra Ames, Dean Snell, Mindy Chase, Kimberly Prescott, and Helen Yu. Thank you! Starting Monday, I’m joining an innovative marketing technology startup. While I can't share all the details just yet, I promise it's a game-changer, and I can't wait to dive in and make an impact. Thank you again for your trust, partnership, and friendship. I'm looking forward to continuing to collaborate, learn from, and grow with all of you in the marketing operations space!

Vinod Khosla (famous VC): “[AI] Agents will act for consumers and fend off marketers and bots.” Well then, now marketers are being equated with bots? Actually, Khosla recently shared a great list of 12 technology predictions for 2035-2049: https://lnkd.in/gx-N-wBa. I pulled the quote out of context to get your attention. Here’s a summary of the AI-related predictions and my assessment of their impact on marketing: 1: Expertise Will Be Near Free The entirety of human knowledge can be stored on a regular phone or computer. It doesn’t just know when the first human set foot on the moon, but also how to create a lead lifecycle or how to organize a webinar. What will happen to today’s experts? Are we the first ones to be replaced by AI? 2: Labor Will Be Near Free This is the physical world equivalent of #1: Khosla predicts that bipedal robots will take care of undesirable jobs. Will we still go to tradeshows and if yes, will the booths be staffed by robots? 3: Computer Use Will Grow Expansively Anyone will be able to do computer programming in human language. We may not even need to do programming anymore, since AI can do the programming for us. Can you imagine telling HubSpot/Marketo what you want and it magically does it for you? 4: AI Will Play a Large Role in Entertainment & Design Music & entertainment will be personalized for you & your mood. Content is created for an audience of one. Not just your favorite TV series, but also nurturing content. 5: Internet Access Will Be Mostly by Agents Agents will act for consumers doing tasks and fending off marketers and bots. This is the one that really gets me thinking: if we - as marketers - have access to AI agents that collaborate to do great marketing, will we still communicate with humans? Or will those humans have their own agents to filter our messages? What would it really mean if agents are doing marketing to other agents? I find this interesting to think about, but maybe also a little scary. What do you think? 

In several recent job interviews, a MOPS manager was asked whether they knew SQL. I’m personally a fan of SQL, but this is what hiring teams should ask instead: “Do you know how to use AI tools to perform relevant MOPS tasks, including creating SQL?” SQL is short for Structured Query Language, the standard way to interact with databases and data warehouses. Data warehouses are getting more important, because they provide richer data for marketing campaigns. So it’s not surprising that some hiring managers expect marketers to know how to query those databases. In today’s competitive job market, it’s important for a job seeker to check as many boxes as possible. But there’s only so much time to upgrade your skills. I’d say, focus on AI, not on SQL. Here’s an example that uses ChatGPT 4 to generate SQL:  1️⃣ Describe your table structure 2️⃣ Describe the desired results  3️⃣ Take the generated query and test it in your database 4️⃣ Tell ChatGPT about any errors or any changes you want to make 5️⃣ Repeat until you’re happy with the results To be able to ask good questions to ChatGTP, it is helpful to know the basics of SQL. If you’re totally new to SQL, here’s a 4-minute primer on SQL: https://lnkd.in/g94eiWRK . If you’re more of a hand-on person, consider doing a 2-hour tutorial with Codecademy: https://lnkd.in/gsQWnciQ . 𝑯𝒆𝒓𝒆’𝒔 𝒕𝒉𝒆 𝒂𝒄𝒕𝒖𝒂𝒍 𝒆𝒙𝒂𝒎𝒑𝒍𝒆: first tell ChatGPT about your database tables, then about the desired results “I have a users table with user_id and email, an orders table with order_id, user_id and order_date, a products table with product_id. product name and unit_price, then an order_product table that links orders and products with the order_id, product_id and quantity. Please write a SQL query for MySQL that returns the user_id, email, and the number of orders for that user in the prior year, plus the name of the most frequently bought product (highest total quantity). It should return only 1 row per user.“ The result is not always perfect, ChatGPT does make mistakes. But it’ll only get better. And if you can do something like this, you certainly have an edge over other candidates. Would you agree learning AI is more important than becoming a SQL expert? 

Are you asked to generate more MQLs because the SDRs are behind on their meeting quotas? It's not you. In his post, Jon Miller explains how the B2B go-to-market has changed, and just cranking up the volume with more emails and more calls isn't the solution. His recommendation is to rethink how you do marketing. Less email, but better email. Properly use AI. Use new signals. Leverage branding and communities. Use new metrics. The challenge is, there are no off-the-shelf products to help with this yet. Today's typical martech products are still stuck in the old paradigm. But that doesn't prevent you from considering the new trends and making gradual changes until a new generation of B2B marketing platforms comes around.

𝐁𝐫𝐢𝐧𝐠 𝐲𝐨𝐮𝐫 𝐨𝐰𝐧 𝐝𝐚𝐭𝐚𝐛𝐚𝐬𝐞. Wouldn’t it be great if all 13,080 Martech vendors would use 𝘺𝘰𝘶𝘳 database instead of each having their own data store? I was talking to Josh Hill about how he was managing his Martech stack when he said “wouldn’t it be great if you could bring your own database to a new Martech vendor”? Often, companies with a large Martech stack already have a central data warehouse with all prospect and customer data. They use this for comprehensive reporting and sometimes also to sync data back to the various apps. But linking up all your apps is a major undertaking. Instead of 𝘺𝘰𝘶 having to put massive efforts into integrating all apps with a data warehouse, shouldn’t it be the responsibility of each vendor to make that connection? There are some challenges of course. How do you standardize the table structure so that a vendor knows where to find leads, contacts, accounts, etcetera? And how to append to it? Which data warehouses would a vendor need to support? Will the data warehouse perform well with so many vendors using it? How do you do identity resolution, meaning, how do you combine data about a single person from various apps? I don’t expect that this will become reality anytime soon, but you gotta have something to dream about. Do you think “bring your own database” could ever happen?