Module 3 — Process design

Module time ~3 hours (reading + exercises) · ~3,500 words · 9 cited primary sources · Tool-independent · v1.0 · 2026-06-09

3.0 Workflows vs processes — the operations layer

Module 2 gave you the five workflows that produce marketing output. This module is about the layer above them: the operating system that decides which workflow runs, when, by whom, and how you know it's done. The distinction matters. A workflow answers "how do we write a brief?" A process answers "how does a request become a brief in the first place, who decides whether it deserves one, what happens when three requests arrive at once, and who signs off before launch?"

Solo founders often believe they can skip this layer — there's only one person, so coordination is free. The belief is wrong for two reasons. First, even a solo marketer coordinates with their own future self: the founder who decides Monday what to ship Friday needs intake, prioritization, and a definition of done, or Friday arrives with three half-finished things. Second, the entire point of Foundations is to build the function so it survives your first hire (the track's Persona B). A marketing function whose process lives in the founder's head cannot be handed to anyone. The process layer is what makes the function transferable — which is the difference between hiring help and hiring chaos.

The intellectual tradition here is operations management, not marketing. Andy Grove's High Output Management (1983) — written about running Intel, adopted as canon by generations of operators — supplies the framing: every recurring activity is a production process with inputs, throughput, quality checks, and output, and the manager's leverage lives in improving the process, not heroically working inside it. W. Edwards Deming's quality tradition supplies the corollary: most failures are failures of the system, not of the people working in it — so when marketing output disappoints, redesign the process before blaming the effort.

3.1 The intake system — how work enters marketing

Marketing is the most interruptible function in a small company. Sales wants a one-pager today, the founder saw a competitor's campaign last night, a partner needs co-branded assets, someone read about a new channel at breakfast. Without an intake system, marketing becomes whatever was requested most recently and most loudly — what operations literature calls the squeaky-wheel allocation, which is indistinguishable from having no strategy at all.

The three components of intake

1. A single front door. All requests — from sales, from the founder, from yourself — land in one place: a shared document, a form, a board column, a dedicated channel. The medium is irrelevant; the singularity is everything. Requests that arrive by hallway conversation get written into the front door by the recipient, or they don't exist. This rule feels bureaucratic for a three-person company and saves the company at five.
2. A triage standard. Each request gets classified within a fixed window (weekly is fine at small scale) into one of four lanes: Do now (genuinely urgent and aligned — rare), Schedule (aligned, goes into the next planning cycle), Clarify (sender gets questions back — most requests are underspecified), or Decline with reason (off-strategy, and the reason is stated, which trains requesters over time).
3. A visible queue. The scheduled lane is publicly visible and ordered. When the founder asks "why isn't X done?", the queue answers: X is third, behind Y and Z, which you agreed were more important. The queue converts prioritization fights from recurring arguments into a single recurring decision.

The triage standard needs one explicit test: does this request serve the current campaign objective (the canvas from Module 2, §2.5) or the standing brand investment (Module 1, §1.6)? Requests that serve neither are not free even when small — every context switch taxes throughput, a finding the operations literature has replicated for decades. Saying no, with the reason, is a process feature.

3.2 Decision rights — DACI for marketing

The second component answers the question that quietly burns the most calendar time in small companies: who actually decides? Marketing decisions — the positioning, the campaign budget, the homepage headline, the discount — tend to attract opinions from everyone and ownership from no one. The result is either decision-by-committee (slow, mediocre) or decision-by-whoever-spoke-last (fast, incoherent).

The standard remedy is a decision-rights framework. Several exist — RACI from project management, RAPID from Bain — but the variant best suited to marketing's speed is DACI, developed inside Intuit in the 1980s and widely popularized in the software industry's playbooks:

The framework's value is not the acronym; it's the two "exactly one" constraints. One Driver means the decision has an engine. One Approver means the decision has an address — and when it's wrong, a learner. For a founder-plus-first-hire marketing function, the standing allocation worth writing down: the hire Drives and Approves asset-level decisions (copy variants, posting schedule, email subject lines); the hire Drives and the founder Approves campaign-level decisions (canvas, budget, channel mix); the founder Drives and Approves strategy-level decisions (positioning, pricing, brand). Writing this down once ends the daily renegotiation of who-decides-what, which is among the largest hidden taxes on a first marketing hire's productivity — and a major reason early marketing hires fail (a theme Module 5 returns to).

3.3 Cadence — sprints, kanban, and the marketing calendar

The third component is rhythm: the recurring schedule on which planning, execution, and review happen. Cadence is what separates a marketing function from marketing episodes. The two dominant patterns come from the software tradition, and marketing borrows both.

Sprint cadence (Scrum-derived)

The Scrum framework — formalized by Ken Schwaber and Jeff Sutherland in the Scrum Guide (first published 2010, refined since) — organizes work into fixed-length cycles with four ceremonies: planning at the start, a short daily sync, a review of output at the end, and a retrospective on the process itself. The "agile marketing" movement adapts this wholesale: a one- or two-week marketing sprint, planned against the campaign canvas, reviewed against the brief's measurement field.

What sprints buy a marketing function is commitment scoping — the discipline of deciding, at the start of each cycle, what will ship by the end, and protecting that commitment from mid-cycle requests (which go to intake, §3.1, not into the sprint). What they cost is flexibility: a two-week commitment is wrong for reactive work like trend-jacking or PR response.

Kanban cadence (flow-based)

The alternative is kanban, adapted from Toyota's production system into knowledge work by David J. Anderson (Kanban: Successful Evolutionary Change for Your Technology Business, 2010). No fixed cycles: work flows continuously through visible stages (queued → in progress → review → done), governed by one mechanism — the work-in-progress (WIP) limit, a hard cap on how many items may be in progress simultaneously. When the cap is reached, nothing new starts until something finishes.

The WIP limit is the single most valuable operations concept for a solo marketer, because the solo marketer's natural failure mode is eight parallel half-finished initiatives. A personal WIP limit of two or three — enforced on a visible board, even a paper one — converts that scatter into a finish-first discipline. Little's Law, the queueing-theory result underneath kanban, formalizes the intuition: for a given throughput, the more items in progress, the longer every item takes. Starting less finishes more.

The hybrid most SMB marketing functions land on

In practice: sprint cadence for campaign work (planned, committed, reviewed on the cycle), kanban with WIP limits for reactive work (a small protected lane for sales requests and timely opportunities), and a quarterly planning checkpoint where the campaign canvas portfolio is set against the year's objectives. Plus the one cadence marketing has that software doesn't: the marketing calendar of immovable external dates — seasonality, industry events, launch windows — which is planned backward from, never toward. The calendar is the skeleton; sprints flesh it out.

3.4 Cross-functional handoffs

The fourth component governs the borders where marketing meets the rest of the company. Handoffs are where work dies: the lead that sales never followed up, the product launch marketing learned about a week out, the invoice-confusing discount finance discovered at month-end. Each border needs an explicit, written contract — three are universal.

Marketing → Sales: the lead handoff

The chronic failure at this border is definitional: marketing celebrates "leads" that sales considers noise. The fix is a jointly written lead definition — the explicit criteria (fit, behavior, expressed intent) that make a contact worth sales' time — plus two service-level agreements: marketing commits to passing only definition-meeting leads with context attached (source, content consumed, stated interest), and sales commits to a follow-up window and to returning disposition data (contacted, qualified, closed, rejected-because). The disposition data is the part everyone skips and the part that matters most: without it, marketing cannot compute which channels produce revenue rather than mere volume — the LTV/CAC discipline from Module 1 §1.6 starves without it. Module 5 treats the relationship side of this border; here, the mechanism.

Product → Marketing: the launch handoff

The chronic failure is timing: marketing discovers the launch when it's too late to do more than announce it. The contract: marketing is Informed at planning, Contributor at naming and positioning, and owns the launch brief (Module 2, §2.1) from a fixed lead time before release — written into the product process, not renegotiated per launch.

Marketing → Finance: the spend handoff

The contract: a monthly budget-versus-actual review on a fixed date, pre-agreed category definitions (what counts as marketing spend vs sales spend vs product spend), and advance notice of any commitment above a stated threshold. Unglamorous, and the foundation of the budget-defense capability Module 5 and Module 7 build on: a marketing function that can't account for its spend can't defend its spend.

3.5 Definition-of-done and pre-launch QA

The fifth component answers "is it finished?" — a question that, unanswered, produces marketing's most public failures: the broken link in the email blast, the typo in the headline, the ad pointing at a dead landing page, the campaign that launched with last year's pricing.

Definition-of-done (DoD)

Borrowed directly from Scrum: a written checklist, agreed in advance, that an asset must pass before it counts as complete. Not aspirational quality standards — binary checks. A marketing DoD for any outbound asset, minimally:

1. Brief conformance — the asset delivers the brief's single proposition to the brief's audience (Module 2, §2.1). Assets that drifted from the brief during production fail here, which is the point.
2. Factual accuracy — every claim, price, date, and name verified against a current source. Claims requiring substantiation flagged per Module 6.
3. Functional integrity — every link clicked, every form submitted, every tracking parameter (UTM) present and correctly formed, on both desktop and mobile.
4. Brand and legal conformance — visual identity correct, required disclosures present (Module 6), consent mechanics intact for email (CAN-SPAM/GDPR basics from Module 4 and 6).
5. Measurement readiness — the brief's deciding metric will actually be observable: tracking verified before launch, not reconstructed after.

Pre-launch QA as a ritual

The DoD is the standard; QA is the enforcement ritual. Atul Gawande's The Checklist Manifesto (2009) supplies the design principles, drawn from aviation and surgery: the checklist is short, binary, performed at a defined pause point, and — critically — performed by someone other than the maker where headcount allows, because makers are systematically blind to their own errors. Solo founders substitute time for headcount: QA happens after a deliberate gap (write Tuesday, QA Wednesday), never in the same sitting as production. The Foundations portfolio includes a pre-launch QA report precisely because this two-hour skill prevents the most expensive category of marketing error: the public one.

3.6 The improvement loop — retros and PDCA

The five components above are the static design. The sixth element makes the design self-correcting. The root pattern is Deming's PDCA cycle — Plan, Do, Check, Act — the foundational loop of the quality movement (Deming, Out of the Crisis, 1986, building on Walter Shewhart's work): plan the change, run it, check the result against expectation, and act on the gap by adjusting the process.

The marketing implementation is the retrospective: a recurring, short, blameless review — end of each sprint or month — with exactly three questions: What worked? What didn't? What one process change do we make next cycle? The constraint to honor is "one change": retros that produce five-item improvement lists produce zero improvements. One process change per cycle, written down, checked at the next retro. Twelve cycles a year compounds into a function that is measurably better run each quarter — and the retro log itself becomes an asset: it is the documented history of why the process is shaped the way it is, which is exactly what your first hire needs to read on day one.

Note what this module did not require: software. Intake can be a shared doc, the queue a numbered list, the board three columns on a whiteboard, the DoD a printed checklist, the retro a calendar recurrence. The tool-options matrix (Module 2, §2.6) applies when volume demands tooling. The process is the capability; the tool is, as ever, optional.