We've Seen That Movie, Part 3: Money, Metrics, and Machinery

Tom Williams, CEO

By the time a movie franchise reaches Part Three, it usually faces a choice: go bigger again, or change the flavor.

We’re choosing the second option.

Parts One and Two stayed largely in the territory of relatively modern, mainstream

Hollywood. This time, we’re moving through the decades: 1930s Charlie Chaplin being swallowed by the machinery of industrial production; 1960s Mel Brooks turning Broadway finance into an elaborate criminal conspiracy; and the 2008 financial crisis, retold through a group of investors watching the world economy collapse while everyone else insists it is probably fine.

Despite the different eras and genres, the underlying theme is consistent. This time, we’re looking at the systems created to support good judgment, and what happens when they begin replacing it: a budget that has to be spent because the financial year is ending; a dashboard that becomes authoritative because the numbers are available, rather than because they matter; a content engine that keeps producing because stopping would feel like failure, even when nobody can clearly explain what the output is achieving.

From the inside, all of this can look reassuringly disciplined: the budget is moving, the numbers are green, and the content calendar is full. But the reality is that the system may be optimizing in entirely the wrong direction.

Part Three is about money, metrics, and the quiet moment when the machinery designed to support the work begins dictating it instead.

The Producers

The Producers is Mel Brooks’s 1967 comedy about Broadway, fraud, and the discovery that it is theoretically possible to make more money from a catastrophic failure than a successful production.

Zero Mostel plays Max Bialystock, a once-successful producer reduced to raising money from wealthy older women for plays that may or may not ever appear. Gene Wilder is Leo Bloom, the anxious accountant who notices something interesting in the books: if a producer sells considerably more than 100% of a show, a flop could generate a fortune because nobody would expect a return.

They therefore set out to produce the least successful musical in Broadway history, overselling shares in it and planning to disappear with the money once it closes.

They find an appalling script, hire the wrong director, cast the wrong leading man, and assemble a production seemingly guaranteed to offend everyone who sees it.

Unfortunately, the audience thinks it is satire.

The show becomes a hit, which is rather inconvenient when your entire business model depends on nobody asking where the money went.

The scheme works perfectly, right up until the production succeeds.

Same Plot, Different Cast

The Developer Marketing Version of The Producers:

When the financial model rewards activity rather than impact, an organization can become remarkably efficient at producing the wrong thing.

Unlike The Producers, technology marketing rarely involves deliberately creating something terrible in the hope that nobody will notice.

That said, the film’s underlying problem is familiar: the people funding the work, producing the work, and evaluating the work are often operating according to a financial logic that has become detached from the outcomes the work is supposed to create.

A program has a budget, a scope, a long list of deliverables. The team is working hard, the production machinery is humming, and everything is being delivered to plan. The only unresolved question is whether any of it has been designed to produce the outcome the business actually needs.

This shows up in a number of ways. Sometimes a budget has to be spent before the end of the quarter, otherwise it will be reduced next time. Sometimes the annual plan was agreed before the product was planned or the market changed. Sometimes the organization already has a well-established events team, content engine, paid media allocation, or agency contract, and those capabilities need to be kept busy.

The financial model quietly becomes the strategy.

A campaign is commissioned because funds are available; an event sponsorship is secured because the budget deadline is approaching; more content is added purely because the production capacity exists.

Budgets are meant to translate priorities into choices, but the incentives surrounding them often point in the opposite direction. Spend less than expected, and next year’s allocation may be reduced. Fail to commit the funds on time, and they disappear. Ask for a pause while the team works out what the audience actually needs, and you may be accused of lacking momentum.

So the organization starts moving: the brief is written around the budget, the scope expands to absorb it, and the measures of success are retrofitted once the activity has already been agreed.

This is particularly dangerous in developer marketing, because the most obvious places to spend money are not always where the real barriers sit. A company may invest heavily in awareness when developers already know the product but cannot understand the documentation. A company might sponsor a major event when the more urgent need is a credible self-service experience. They might commission more content when the positioning is still too vague for any of it to say something useful to the target audience.

The resulting program can look substantial, with deliverables, media plans, production schedules, and a reassuringly depleted budget line.

But a full production schedule is not evidence that the right production has been funded.

Before committing the investment, it helps to ask what needs to change. Is the problem awareness, understanding, evaluation, onboarding, integration, or sustained use? Which audience matters most? What is preventing them from progressing? Is marketing even the function best placed to fix it?

A budget can fund an answer. It cannot tell you what the question should have been.

Changing the Ending

The answer lies in creating enough space between receiving the budget and activating it to make a deliberate choice.

This is where Catchy can make a real difference. Typically in such cases, we use a small part of the available budget to build the map first: use our Developer Voice social listening methodology to understand which topics are already gaining traction; run a stakeholder workshop to prioritize audiences; conduct a review of the developer journey to identify where people are getting stuck; or run an audit of existing content to understand what is working, what is missing, and what no longer deserves further investment.

None of this needs to become a lengthy strategy exercise. The purpose is to create enough clarity for the rest of the budget to be spent intelligently.

That may still lead to a campaign, an event, a content program, or paid media. It may also reveal that the highest-value intervention sits in documentation, onboarding, research, or product experience.

The point is not to spend less. It is to build the map before committing to the route.

Before you fund the production, it helps to agree what would count as a hit.

The Big Short

The Big Short is a film about the global financial crisis, which does not sound like an especially entertaining way to spend two hours. Somehow, it is.

It follows a group of investors who realize that the American housing market is built on loans far weaker than the financial system is prepared to admit. The mortgages are failing, the models are wrong, and the supposedly safe products assembled from them are being held together by confidence, jargon, and the belief that risk disappears once it has been divided into enough acronyms.

The evidence is there. The difficulty is that the entire system has been organized around not seeing it.

Banks are making money, ratings agencies are approving the products, and the numbers continue to provide a comforting account of reality long after reality has quietly left the building. The people who notice the gap are treated as eccentric, irresponsible, or simply too early, which in finance is often another way of saying correct before it became socially convenient.

There are several striking moments where everyone looks at the same information and reaches the conclusion most compatible with their bonus structure.

The numbers are not entirely fictional. They are simply being asked to support a story they can no longer justify.

Same Plot, Different Cast

The Developer Marketing Version of The Big Short:

Organizations often measure what is available rather than what matters.

Marketing teams today have access to more data than ever: impressions, clicks, views, registrations, downloads, followers, engagement rates, page visits, form fills, leads, and any number of variations produced by the platforms distributing the work.

The difficulty is not usually a lack of numbers. It is working out which ones deserve authority.

John Doerr’s Measure What Matters has become the obvious shorthand for this problem, and the title alone contains much of the lesson. Measurement is not valuable simply because it is precise or attractively presented. It’s valuable when it helps us understand something that actually matters.

A metric becomes influential partly because it is easy to capture. Social platforms provide engagement data, so engagement enters the report. Event systems provide registrations, so registrations become evidence of demand. A dashboard displays traffic in a pleasing upward curve, and the curve begins to stand in for progress.

None of these numbers is inherently meaningless. The problem begins when a measure of activity is allowed to answer a question about impact.

A campaign can generate impressions without increasing understanding; a report can be downloaded by people with no relationship to the buying journey; a developer can create an account, fail to complete the first meaningful action, and still appear in the conversion numbers as a success.

The metric is accurate, but the interpretation is not.

The outcomes that matter are usually slower and harder to observe. Adoption may involve several people, multiple systems, and a long delay between first contact and meaningful product behavior. A user may discover the product through one channel, evaluate it through another, and require approval from colleagues who never encounter the original campaign at all.

By comparison, a click is wonderfully cooperative: it appears immediately, fits neatly into a dashboard, and rarely asks difficult questions about causation.

Large organizations can add further complication because different teams own different parts of the evidence. Marketing can see campaign performance, product can see usage, sales can see opportunities, and customer teams can see retention. Each group has a partial view, and each is rewarded against measures it can influence directly.

The result is often a collection of individually reasonable numbers that collectively fail to describe the whole system.

This is where the analogy with The Big Short becomes useful. The financial crisis was not caused by an absence of information. It was enabled by incentives, assumptions, and models that made contradictory evidence easier to dismiss than confront.

Developer marketing has lower stakes, thankfully, but the pattern is familiar: the dashboard says that engagement is up and the campaign has delivered, while actual product usage remains flat, or developers arriving at the documentation fail to complete a first run experience.

The purpose of measurement is not to prove that the work happened. It is to improve our understanding of whether the work changed anything.

Changing the Ending

My advice is consistently to start with the decision the measurement needs to support. If the objective is awareness, reach and recall may matter. If it is evaluation, look at movement into documentation, trials, or meaningful product exploration. If it is adoption, our measures need to reach further into activation, integration, repeat use, or expansion.

No single dashboard will explain everything, but the evidence should at least connect the activity to the behavior it was intended to influence.

This is another way Catchy can help: connecting campaign, product, and commercial evidence; testing whether current metrics answer the questions that matter; and building a measurement framework that follows the audience from first contact to meaningful action.

That does not mean pretending every outcome can be neatly attributed. It means being honest about what the data can prove, where the gaps are, and which additional signals would make the picture more useful.

The answer is not to distrust numbers. It is to ask what story they are capable of supporting.

Above all else, remember that numbers may be perfectly accurate while still answering the wrong question.

Modern Times

Modern Times was released in 1936, almost a century before content calendars, marketing automation, and the instruction to “do more with less.” Charlie Chaplin appears to have anticipated all three.

Chaplin plays a factory worker whose job is to tighten bolts as they pass him on an assembly line. The machinery speeds up, the work becomes more relentless, and every human need is treated as an obstacle to greater efficiency.

At one point, he is volunteered to work with a machine designed to feed workers while they continue working, because lunch has apparently become an unacceptable drain on productivity. The demo goes badly. The film’s most iconic image has Chaplin pulled inside the machinery, moving helplessly through a system he is supposed to be operating.

That’s the joke, and the warning: the machines were built to make the work more efficient; instead, the machine begins determining what the work, and the worker, are for.

The production line never asks whether what it produces is useful. Its job is simply to keep moving.

Same Plot, Different Cast

The Developer Marketing Version of Modern Times:

The machinery created to support marketing can end up dictating the work.

Most marketing systems begin with sensible intentions. Content calendars, workflows, templates, automation, and reporting are all designed to make the work more consistent, accountable, and scalable, allowing teams to deliver programs that would be difficult to manage through improvisation alone.

The problem begins when the system acquires momentum of its own.

An article is due because the calendar says an article is due. The newsletter goes out because it always goes out. Social posts are produced to fill the channel. Events are added because the annual plan contains a space for them. The work continues, not because each piece has a clear purpose, but because interrupting the machinery would feel like failure.

Once an organization has assembled a content engine, reducing the output can look like dismantling capability. The team exists, the budget has been allocated, the workflow is established, and the reporting cadence expects something to measure. It becomes easier to produce another asset than to ask whether the asset deserves to exist.

The machinery also begins shaping the work itself. Templates encourage the same formats. Approval processes reward the safest language. Search requirements pull every article toward the same topics. Performance systems favor what has worked before, even when what worked before is no longer what the audience needs.

The result is often efficient sameness.

AI makes this both more powerful and more dangerous. The cost of producing content is falling, but the cost of weak judgment is not. A poor assumption can now travel through the system faster, appearing as ten articles, twenty social posts, three email variants, and a set of campaign assets before anyone has stopped to ask whether the original idea was worth scaling.

More output is only useful when the organization knows what deserves to be produced.

Without that clarity, people begin serving the calendar, feeding the channels, and supplying the algorithm. The process that was meant to support good work becomes the reason the work exists.

This does not make process the enemy. Developer marketing is too complex to run on inspiration alone. Teams need systems, especially when they are coordinating across products, regions, channels, and internal stakeholders.

But those systems should extend judgment, not replace it.

A good production model makes useful work easier to create, improves its quality, and helps the organization learn. A bad one simply increases the speed and consistency with which the wrong work is produced.

The moment the calendar determines the strategy, the system is no longer supporting the work. The work is supporting the system.

Changing the Ending

The single best move here is to create deliberate points where the marketing machinery has to justify itself.

What is this content meant to change? Who needs it? Why this format? Why now? What would happen if we did not produce it?

Those questions should not become another elaborate approval process. Their purpose is to preserve the connection between output and intent.

Teams also need permission to stop: retire formats that no longer serve the audience, reduce cadence when quality or relevance is falling, and use automation where it removes low-value effort while preserving human judgment where context, originality, and restraint matter.

At Catchy, this often means reviewing the production system as well as the output: which parts are creating value, which are merely consuming capacity, and where a lighter process would produce better work.

A production system should make useful work easier. It should not make unnecessary work inevitable.

The machine is meant to serve the strategy, not consume it.

We've Seen That Movie

Money, metrics, and machinery all begin as tools: a financial model funds the work, measurement provides evidence, and a production system turns plans into repeatable output.

The trouble begins when the tool develops a logic of its own.

In The Producers, the entire production is designed around a financial outcome that has almost nothing to do with the quality or success of the work. In The Big Short, reassuring numbers protect a version of reality that is already collapsing. In Modern Times, the machinery created to improve production gradually takes control of the person operating it.

Developer marketing has its own, generally less catastrophic, versions of all three: programs designed around the available budget rather than the problem; dashboards reporting success without showing whether behavior changed; content engines producing work because the engine needs to be fed.

None of this means money, metrics, or operational systems are the enemy. It means they are poor substitutes for judgment.

A budget cannot decide where it will have the greatest effect. Data cannot determine which question matters. Machinery cannot tell us whether the thing it is producing deserves to exist.

That responsibility remains stubbornly human.

Part Three has been about the systems surrounding the work. In Part Four, we turn to the people inside them: the different roles involved in adoption, the many identities organizations attempt to squeeze into DevRel, and the point at which a gap between promise and reality becomes a question of trust.

An ensemble cast, several versions of the same person, and a whistleblower.

What could possibly go wrong?

As always, if any of these plots sound familiar, we’re always up for a chat.