ROB YOUNG'S PIPE DREAMS PODCAST 

Stop Scaling the Wrong Things

Rewriting the early-stage marketing job description for the AI era

Host: Rob Young, MarketingHQ     Guest: Mike Harris, Harris CMO Partners

The premise

A key insight from a recent dinner meeting is that AI makes it much easier to scale sales and marketing, which means it has never been easier to scale the wrong things. The first instinct is to use AI to increase output. The bigger opportunity is to use it to improve the inputs, and to build marketing that moves deals forward and supports renewals instead of just handing over leads.

Most early-stage marketing manager job descriptions haven’t caught up. They still hire for output: channels, content volume, MQLs, and seldom mention sales pipes. This episode rewrites one, live, piece by piece, so founders and marketing managers leave with a concrete picture of what the job should look like now.

Podcast Discussion Guide

1. Scaling the wrong things   ~4 min

ROB ASKS

Mike, you’ve been talking about this idea that it’s never been easier to scale sales and marketing with AI, which means it’s never been easier to scale the wrong things. Where did that come from?

MIKE’S TALKING POINTS

• AI has made much of marketing’s output nearly free but most marketing manager job descriptions still focus on output. That mismatch is the problem.

• The pressure to prove pipeline conversion is here and will stay. Investors are running tougher math: Bain estimates a typical buyout now needs 10–12% annual EBITDA growth to earn the return that 5% delivered a decade ago. Venture investors say the same thing with different words: burn multiple or CAC payback for example.

• If a job description describes output, AI will make the role ornamental. If it describes judgment and outcomes, and is linked to the sales pipe, AI makes the role more valuable.

So what happens to a marketing role that’s still written around output?

• It gets changed, and rewritten for several things. These are the 2 most important I’m seeing: creating systems that create the output, and KPIs related to sales pipe impact get added in.

CLIP LINE

“At 5%, marketing could be judged on reach. At 12%, it gets judged on revenue.”

2. The “before” job description   ~4 min

ROB ASKS

What does a typical early-stage marketing manager job description look like today, and what is wrong with it?

MIKE’S TALKING POINTS

• It’s often borrowed from a later-stage company with lots of channels, a current list of activities, and an MQL target.

• These marketing job descriptions are often not written by marketing experts, and not customized for a company’s or founder’s current situation. 

• For example, “Familiarity with AI tools a plus” sits at the bottom as a nice-to-have, when it should shape the whole role.

• It rarely names a budget, a pipeline number, or who makes which decisions.

EARLY-STAGE REALITY CHECK

What about companies that haven’t found product-market fit yet? Should the job description even look the same?

• Before product-market fit, the job is to find, not to scale.

• However, discovering product-market fit is a lot like finding prospects. AI can only take you so far, and you would never let AI make the decision on product market fit. That’s a judgment call. But you use the same marketing techniques to reach out to potential prospects and test your PMF with them. Start with creating some social media posts, maybe talking about an anonymized version of what you’re thinking about. Ask people to weigh in…everyone has an opinion. So you are quietly building pre-demand before the release of your product.

• This suggests that job descriptions might show phases over the near and intermediate terms. Near term focuses on solidifying PMF and quietly testing your idea with your ideal client profile. Once you have that you’ll more than enough to start building real demand with marketing.

JOB DESCRIPTION EDIT

We put the composite “before” version on screen (see the appendix).

3. Inputs over outputs   ~6 min

ROB ASKS

Most teams’ first instinct with AI is to make more stuff. More content, more emails, more campaigns. You say that’s the wrong place to start. Why?

• More outputs delivered to the same market reach will not deliver more conversions.

• Better outputs created with better inputs delivered to the same or expanded market reach will deliver more conversions.

But at most startups the data is pretty thin. A handful of deals, messy CRM. How do you make decisions when you can’t fully trust the numbers?

• Again, you talk to people. One of the best books I ever read on this told the story of a founder who talked to hundreds of prospects before ever finalizing product design. That used to be done with telephone calls but today we have social media and AI to help us find and engage our ideal client profile.

MIKE’S TALKING POINTS

• One key to creating better inputs is to use AI to analyze the current hot topics in a market space and to find and analyze conversations around those. 

• Also consider creating a pipeline based on those conversations. Again, another great opportunity to practice something that will soon be critical for you. We can use AI to analyze what happens, in call recordings and notes, emails, social media responses. These can include won, lost and stalled deals, objections, churn reasons, and ICP fit. Use the data to help you decide what marketing to create.

EARLY-STAGE REALITY CHECK

However you should question the numbers in AI responses, because they don't always hold up. AI can and should flag when a sample is too small to trust, so ask it to show confidence levels. What it can't tell you is whether the data measures the right thing (inconsistent CRM stages, skewed attribution, a dozen deals from the founder's network), and at an early stage you still have to decide before the numbers are conclusive.

JOB DESCRIPTION EDIT

Responsibilities change. “Produce eight blog posts a month” becomes “Run a monthly review of won, lost and stalled deals and use it to set the content and campaign plan.”

4. Beyond the handoff   ~7 min

ROB ASKS

If marketing should do more than hand over leads, and help move deals forward and support renewals, what should a marketing manager be measured on?

MIKE’S TALKING POINTS

• First of all, replace the MQL count with sales-ready handoffs: leads that already meet an agreed share of sales’ own qualification criteria, with the evidence attached. This means a marketer’s job description now tilts heavily toward interaction with prospects before sales even knows those prospects are there.

• For example, take a typical eight-point SQL checklist. Marketing can own two of the criteria outright: revealed pain points and active evaluation of competitors. Plus it can gather strong signals on four more: purchase criteria, user expectations, timing and fit with their current systems. Well crafted quizzes presented at the right time are great for this. Qualifications like budget and organizational expectations stay with sales because, in my opinion, those are best done in live conversations.

• AI makes this practical. It can help assess Intent data, technographics (which is data about a company’s tech stack), answers from interactive assessments, ROI calculators and chatbots on the website all contain signals marketing managers cannot track by hand.

• Then follow the deal: stage-to-stage conversion, sales-cycle length, win rate in the target segment, and marketing’s contribution to each of these.

• Some employers already write this in. For instance, one current product marketing posting at a Sequoia-backed startup says the role evolves “from hand-crafting every asset to designing the system that produces them.”

• So we’re seeing the marketer’s time and activity evolving from producing assets to deciding which assets are worth producing. That’s the most important piece of content strategy.

THE FUNNEL MATH

Working back from a goal of 500 SQLs, with planning benchmarks of 2% reach-to-conversion, 25% conversion-to-MQL and 13% MQL-to-SQL, marketing needs to reach about 770,000 people. If pre-qualifying against the SQL criteria raised MQL-to-SQL to 20%, the required reach would fall to about 500,000, roughly a third less, for the same pipeline.

CLIP LINE

“Better inputs beat more output: a third less reach for the same pipeline.”

EARLY-STAGE REALITY CHECK

Smaller teams, tighter budgets: at early stage, sales time is the scarcest resource, and every unqualified handoff spends it. But don’t turn forms into interrogations. Asking about budget on a form reliably hurts conversion. Gather signals; let sales confirm.

QUESTION FOR THE FOUNDER LISTENING

Will you give marketing full visibility into CRM and pipeline data, and can sales come to an agreement on the SQL criteria, and stick with it for a specified period of time? You cannot hold someone to a number they cannot see.

JOB DESCRIPTION EDIT

The metrics section replaces the MQL target with sales-ready handoffs against agreed SQL criteria, then follows deals through close and renewal.

5. Founder dynamics   ~5 min

ROB ASKS

At early stage the founder is often the best marketer in the building and the final approver. How should that shape the job?

MIKE’S TALKING POINTS

• The problem is that most job descriptions, and even interviews, are silent on who decides what. That silence is where first marketing hires can fail.

• One way to address this is to name the decision rights up front: what the marketing manager owns, what the founder approves, what is shared.

• Another way is to treat the founder’s voice as an asset to deploy, with marketing using AI to help create a founder voice engine, which can be referenced for all related content creation. This takes some of the marketing burden off the founder’s shoulders and puts more control in marketing’s hands.

• Watch for the “Lead” title with an individual-contributor mandate. Say which one it is.

QUESTION FOR THE MARKETING MANAGER LISTENING

How do you find out in the interview how much the founder will really let go?

JOB DESCRIPTION EDIT

A new “Decision rights” section.

6. What the team looks like   ~5 min

ROB ASKS

What should an early-stage marketing team look like now that AI can do much of the production work?

MIKE’S TALKING POINTS

• A marketing manager, a deliberate AI toolset, and fractional specialists can cover what used to take three or four hires.

• Before hiring a second person, automate the production work and see where the real bottleneck is.

• The second hire follows the bottleneck: product marketing if positioning and fit are unresolved, demand generation if the motion is proven and needs volume.

EARLY-STAGE REALITY CHECK

Smaller teams and tighter budgets: the job description should say what the person actually has to work with.

JOB DESCRIPTION EDIT

A new “Resources” section: budget, tools, data access, fractional support.

7. The AI roadmap   ~5 min

ROB ASKS

Mike, you said rethinking job descriptions needs a roadmap for AI to accelerate efficiency and conversion. What goes in it?

MIKE’S TALKING POINTS

• Days 1–30, audit: map the funnel, pull two quarters of won and lost deals, talk to customers, set up AI-assisted analysis of calls and CRM data, set baselines.

• Days 31–60, rebuild the inputs: sharpen the ICP and messaging from what the data shows, build repeatable AI workflows for routine production, run one experiment aimed at a specific pipeline stage.

• Days 61–90, prove it: report the change in conversions, stop what isn’t working, propose next quarter’s plan and the next hire.

EARLY-STAGE REALITY CHECK

No runway to waste: a 90-day outcome replaces the list of responsibilities as the heart of the job description.

JOB DESCRIPTION EDIT

A “First 90 days” section completes the “after” version.

8. Close   ~3 min

For founders: Would your current marketing job description survive being graded on pipeline?

For marketing managers: If AI did half your job tomorrow, which half would it be, and what is left?

The marketing manager job description, before and after

The “before” is a composite of typical early-stage postings, not any one company’s. Bracketed items are placeholders a founder would fill in.