Every month, the same meeting happens in growing software companies across South Africa. Marketing presents a chart showing more qualified leads than last month. Sales says most of those leads were not worth a phone call. Someone suggests a new tool, a new campaign or a new hire. Nobody can prove who is right, because the two teams are counting different things in different systems, and the information that would settle the argument never travels back from the people who spoke to the customer.
That is not a people problem, and it is rarely a tool problem. CloudNala's view is that a go-to-market motion is a system, and it behaves like one whether or not anyone designed it. The revenue a company wins is decided less by how hard each team works than by whether the stages are defined, whether each has an owner, and whether what later stages learn flows back to earlier ones. Design those three things and most of the arguments go away. Leave them to chance and every new tool makes the noise louder.
This ten-part series treats go-to-market the way an engineer treats any other production system: with written definitions, trustworthy data, testable hypotheses, explainable decisions, small reversible changes and governance. This first part sets out the system view. Later parts cover the GTM engineer role, revenue data, signals, lead scoring, plays, AI personalisation, handoffs, measurement and governance.
A note on what this series is. It describes a method, with a fictional company carrying the examples. It is not a report of results from any CloudNala engagement, and the numbers are there to make the mechanics concrete, not to promise outcomes.
Lowveld Ledger is fictional, and its systems and figures are illustrative. Lowveld Ledger is a Mbombela company that sells cloud accounting and payroll software to South African SMMEs. It has about 2,300 paying customers, a 30-day free trial, monthly webinars, a referral channel of roughly 140 accounting practices, and outbound email and LinkedIn work done by two SDRs. Its sales and marketing team of nine is about to hire its first GTM engineer.
The argument that starts every month
At Lowveld Ledger the monthly number looked like this. Marketing reported around 380 marketing-qualified leads a month. Sales said fewer than 90 were worth a call. Marketing replied that sales was not following up. Sales replied that marketing was counting webinar sign-ups from students and competitors.
Both were partly right, and neither could prove it. "Qualified" had never been written down. The CRM held about 41,000 contact records, and nobody knew how many were duplicates. Trial activity lived in the product database, which sales could not see. When an account executive rejected a lead, the reason went into a free-text field, and the most common entry was "not interested".
Each team was doing its job as it understood it. The trouble was that nobody owned the whole.
A system is stages, owners and loops
A useful way to see any go-to-market motion is in three layers.
Stages are the states a buyer moves through: attracted, qualified, converted, onboarded, expanded. The names vary by company. What matters is that each stage has a written entry rule, so that two people looking at the same record put it in the same place.
Owners are the people accountable for each stage and for the movement between stages. A stage with two owners has none. A handover with no owner is where records go to die.
Loops are the flows of information backwards: from sales to marketing about why leads were rejected, from onboarding to sales about which customers actually started using the product, from customer success to everyone about who stayed and grew. Loops are how the system learns.
Most organisation charts and most CRM dashboards show only the stages. That is why most improvement efforts focus on stages too: more leads at the top, faster follow-up in the middle, bigger discounts at the end. Those efforts are not useless, but they are local. They make one number go up without telling you whether the company is attracting the customers it can keep.
Why the broken loops matter more than the weak stages
When Lowveld Ledger's leadership looked at its motion this way, the stages were mostly fine. Marketing produced plenty of interest. The SDRs were busy. The account executives closed deals. Customer success kept most customers.
The loops were where things broke.
Rejection reasons were never usable, so marketing kept targeting the same people. Won accounts were never compared with the company's own idea of a good customer, so the target profile was a guess from three years earlier. Payroll activation, the moment a trial user runs their first real payroll, was the best predictor of a customer who stays, and it sat in a database the sales team had never seen. Exit surveys from churned customers were collected and not read. And accountant partners, the source of the company's best leads, heard nothing back about their referrals, because their status lived in the partner manager's spreadsheet.
This pattern is common because loops cross team boundaries. A stage sits inside one team, so someone notices when it fails. A loop runs from one team to another, so its failure is nobody's job.
The consequence is a system that cannot learn. It can work harder, and it can buy more tools, but it keeps making the same targeting and timing decisions because the evidence that would change them never arrives.
Map your motion before you change it
The first practical step is to draw the system as it really is, not as the sales playbook says it is. Lowveld Ledger did this in a two-hour working session with the head of sales, the marketing lead, the partner manager and the customer success lead in the room.
The map answered five questions for each stage:
- What is the entry rule? If the answer was "it depends", that was written down as a gap.
- Who owns it? One named role, not a team.
- Which system records it? Where would you look to prove a record is in this stage?
- What does this stage learn that an earlier stage needs? This defines the loops.
- How does that learning travel today? Usually the honest answer was "a conversation, sometimes".
The session produced a single page with the stages, owners and systems across the top, and the loops drawn underneath. Seven loops were named. Two worked. Five were broken in one of four ways: never captured, captured as free text, captured in a system nobody else could see, or captured and never reviewed.
That page became the backlog for everything that followed.
Where tools and AI fit
It is tempting to read this and conclude that the answer is a better CRM, an intent-data subscription or an AI agent that writes outreach. Each of those can help, and later parts of this series cover them. But a tool attached to a system with undefined stages and broken loops automates the confusion.
An AI email writer that drafts messages for leads nobody has defined as qualified will produce more, faster, of the wrong thing. A lead score built on a CRM where one in four records is a duplicate will rank duplicates. An intent feed that flags companies "researching payroll software" adds a signal, but nobody can tell whether it predicts anything until the outcome data flows back.
The CloudNala article on why strong AI agents start vertical before going horizontal makes a similar point about agents in general: narrow, well-defined work first. The same principle applies here. Define and instrument the system, then add automation where the system is already sound.
What changes when you treat it as a system
Treating go-to-market as a system changes four habits.
Definitions become documents. "Qualified" stops being a feeling and becomes a rule that is written down, versioned and visible in the CRM. Part 3 covers this in detail.
Every decision rule becomes inspectable. Lead scores, routing rules and play triggers become things a seller can read and challenge, rather than black boxes. Part 5 covers scoring.
Changes become experiments. A new campaign, signal or message is treated as a hypothesis with a test and a decision date, not as a permanent addition. Parts 4, 6 and 9 cover signals, plays and learning.
The loops get owners. Every backward flow of information has a named owner and a place where it lands. Part 8 covers handoffs.
None of this needs a large team. Lowveld Ledger is a ten-person commercial function. The point of the system view is that small teams cannot afford to waste effort on the wrong targets, and they are the teams least able to absorb the cost of a confused CRM. CloudNala has argued before that small teams need better decision systems, not more tools, and go-to-market is one of the clearest cases.
What Lowveld Ledger decided
After the mapping session, Lowveld Ledger made three decisions.
First, the monthly argument would stop until there were written definitions for each stage. Both marketing and sales would report against the same definitions, from the same data.
Second, the five broken loops were ranked by how much decision value they carried. Payroll activation came first, because it was the strongest available evidence of a good customer and it was already being recorded. Rejection reasons came second.
Third, the company would hire a GTM engineer, a role it had seen described in job ads and on LinkedIn without being sure what it meant. Before writing the job ad, the leadership team agreed to define the role by its work products rather than by its title. Part 2 explains what that definition looked like.
Start this week
- Draw your motion on one page. Stages across the top, an owner and a system of record under each, and the backward loops underneath. Do it with the people who run each stage in the room.
- Mark every loop as working or broken, and for each broken one, write which of the four failure types it is: never captured, free text, invisible system or never reviewed.
- Pick the one loop with the most decision value and give it a named owner this week, even before you have the tooling to automate it.
Next: Part 2, A GTM engineer is not a more technical SDR. All parts are on the series page.
How CloudNala can help
CloudNala runs go-to-market systems reviews for growing technology and services companies. We map your motion as stages, owners, systems and feedback loops, identify the broken loops and the definitions that are missing, and turn that into a short, ranked backlog of changes to data, rules and handoffs. We start from the system you have, and we will tell you when the answer is a written definition rather than a new tool.
Work with CloudNala
CloudNala helps organisations move from technology ambition to practical execution across cloud, AI, data, platform engineering and digital services.
Whether you are exploring AI, modernising your cloud environment, building a public-sector digital service, or turning an idea into a working MVP, we can help you shape the roadmap and deliver the next step.
Book a Go-to-Market Systems Review or write to us at consult@cloudnala.co.za