Measuring Marketing

This closing post covers measuring marketing — attribution, marketing metrics, the brand-vs-performance measurement divide, and the mindset for measuring wisely. It ties the series together, because knowing whether your marketing works (and which parts) is what lets you improve it. But marketing measurement has genuine limits that a naive “measure everything” approach misses — and understanding those limits is as important as the metrics themselves.

Why measuring marketing is hard

Engineers expect measurement to be straightforward — instrument it, track it, optimize it. Marketing measurement is genuinely harder, and understanding why prevents both naive over-confidence and giving up:

• Attribution is genuinely difficult. Attribution — determining which marketing activities caused a given customer/sale — is hard because customer journeys are complex: someone might see your brand content, read an article months later, hear a recommendation, then search and convert. Which of those “caused” the sale? Attributing credit across many touchpoints over time is genuinely difficult, and simple attribution (crediting the last click) often misleads by ignoring everything that came before.
• The most valuable marketing is the hardest to measure. Here’s the crux: brand, demand creation, and word-of-mouth — often the most valuable marketing (from the brand and demand-gen posts) — are precisely the hardest to measure (diffuse, long-term, hard to attribute). Meanwhile, easily-measured activities (a specific ad’s clicks) aren’t necessarily the most valuable. So measurability and value are not aligned — which is the central trap.
• Measurability bias distorts investment. The danger, especially for metrics-minded people, is investing in what’s measurable rather than what’s valuable — pouring money into trackable performance marketing while neglecting hard-to-measure brand and demand creation (the brand-vs-demand tension, again). Optimizing only what you can measure can quietly starve the most valuable, harder-to-measure marketing. This bias is the single most important thing to guard against in marketing measurement.

Marketing measurement is hard because attribution is genuinely difficult and — critically — the most valuable marketing is the hardest to measure while measurable activities aren’t always the most valuable. The engineer’s “measure everything” instinct, applied naively, leads straight into measurability bias (over-investing in the trackable). Good marketing measurement navigates this rather than pretending measurement is clean.

Marketing metrics

Despite the difficulty, there are useful marketing metrics — the key is knowing what they tell you and their limits. The important ones connect to the whole series:

• Funnel and conversion metrics — how people flow through the funnel (from the demand-gen post): traffic, leads, conversion rates between stages, and where drop-off happens. These reveal where the funnel is working or leaking and are genuinely actionable (fix the leakiest stage). Useful and relatively measurable, though they capture the funnel, not brand.
• Unit-economics metrics: CAC and LTV — the cost to acquire a customer (CAC) vs their lifetime value (LTV), from the GTM and finance series. These are the most important marketing metrics for viability: marketing must acquire customers for meaningfully less than they’re worth (CAC < LTV, comfortably). CAC/LTV is the bottom-line test of whether marketing is economically working, cutting through vanity metrics.
• Channel performance — which channels produce customers (not just clicks) and at what cost, letting you invest in what works (from the GTM channels post). Measure channels by customers (and profitable ones), not by top-of-funnel activity.
• Beware vanity metrics. Metrics that look impressive but don’t connect to real outcomes — raw traffic, impressions, followers, likes — are vanity metrics (the GTM measurement post): they can rise without the business improving. The discipline is measuring what connects to customers, revenue, and retention, not what merely looks good. Ask of any metric: does moving it actually move the business?

The useful marketing metrics — funnel/conversion (where you win and lose), CAC/LTV (the viability test), and channel performance (what actually produces customers) — are those that connect to real business outcomes, while vanity metrics (traffic, impressions, followers) mislead. Knowing which metrics matter (and resisting vanity ones) is core to measuring marketing well. But even good metrics don’t capture everything, which is the brand-vs-performance divide.

Brand vs performance measurement

A fundamental divide in marketing measurement mirrors the brand-vs-demand distinction: performance marketing (measurable) vs brand marketing (hard to measure):

• Performance marketing is measurable and short-term. Performance (direct-response) marketing — ads and campaigns aimed at immediate, trackable conversions — can be measured relatively directly (spend in, conversions out, calculate ROI). This measurability makes it attractive and manageable: you can optimize it with data. It captures existing demand efficiently and answers “did this campaign produce customers?”
• Brand marketing is valuable but hard to measure. Brand marketing — building long-term reputation and trust — resists direct measurement (its payoff is diffuse, delayed, and hard to attribute), but it’s often more valuable long-term (it creates future demand and makes performance marketing more effective, from the brand post). You can gauge brand indirectly (awareness surveys, branded search volume, share of voice, direct/organic traffic trends), but never with performance marketing’s precision.
• The trap and the discipline. The trap: because performance is measurable and brand isn’t, teams over-invest in measurable performance marketing and under-invest in brand — optimizing the measurable while starving the valuable. The discipline: invest in both, accepting that brand’s value is real even though it can’t be precisely measured, and using indirect brand indicators rather than abandoning brand for lack of clean metrics. Don’t let “we can’t measure it precisely” become “we won’t invest in it” — that’s measurability bias making your decisions.

The brand-vs-performance measurement divide is the crux of measuring marketing wisely: performance marketing is precisely measurable and short-term, brand marketing is imprecisely measurable but often more valuable long-term, and the discipline is investing in both — resisting the pull to fund only what’s cleanly measurable. This is the same brand/demand balance from earlier posts, now framed as a measurement challenge.

Measuring marketing wisely

Bringing the series together, the mindset for measuring marketing well:

• Measure what you can, value what you can’t. Rigorously measure the measurable (funnel, CAC/LTV, channel performance) and use it to improve — and invest in the valuable-but-hard-to-measure (brand, demand creation, word-of-mouth) even without clean metrics, using indirect indicators where possible. The goal is good judgment informed by measurement, not measurement-only decisions. Both the measurable and the unmeasurable-but-valuable get investment.
• Anchor on CAC/LTV and real outcomes. The bottom-line test of marketing is economic: are you acquiring customers profitably (CAC comfortably below LTV), and growing? Anchor on metrics that connect to real business outcomes (customers, revenue, retention), and treat vanity metrics with suspicion. This keeps measurement honest.
• Beware measurability bias above all. The single biggest measurement mistake is letting measurability drive investment — funding trackable performance/demand-capture while starving hard-to-measure brand and demand-creation. Consciously guard against this: the most valuable marketing is often the hardest to measure, so don’t let the measurable crowd out the valuable. This warning has recurred throughout the series because it’s the central pitfall for metrics-minded (engineering) teams.
• Use measurement to improve, iteratively. As with growth marketing, use measurement to find what’s working and what isn’t, and iterate — improving the funnel, reallocating to better channels, testing and learning. Measurement’s purpose is improvement, turning marketing into an iterative, learning system rather than a set-and-forget spend.

Measuring marketing is genuinely hard — attribution is difficult, and the most valuable marketing (brand, demand creation, word-of-mouth) is the hardest to measure while measurable activities aren’t always the most valuable — so the discipline is to measure what you can (funnel, CAC/LTV, channels), value what you can’t (brand), anchor on real economic outcomes, and above all resist measurability bias. That completes the series: from what marketing honestly is, through brand, product marketing, content/SEO, demand gen, developer marketing, and growth, to measuring it wisely. Marketing, done well and honestly, is how good technical work reaches the people it can help.

Key takeaways

• Marketing measurement is genuinely hard: attribution (which activity caused a sale) is difficult across complex multi-touchpoint journeys, and — critically — the most valuable marketing (brand, demand creation, word-of-mouth) is the hardest to measure while easily-measured activities aren’t always the most valuable, so measurability and value are not aligned.
• The useful metrics connect to real outcomes: funnel/conversion metrics (where you win and lose customers), CAC vs LTV (the viability test — acquire customers for comfortably less than they’re worth), and channel performance (what produces actual customers, not clicks) — while vanity metrics (traffic, impressions, followers) look impressive but can rise without the business improving.
• Performance marketing (direct-response, immediate, precisely measurable) vs brand marketing (long-term reputation, valuable but only indirectly measurable via awareness/branded-search/organic trends) is the measurement crux — and the trap is over-investing in measurable performance while starving harder-to-measure but often more-valuable brand.
• The single biggest measurement mistake is measurability bias — letting what’s measurable drive investment, funding trackable activities while neglecting hard-to-measure brand and demand creation — a recurring pitfall for metrics-minded engineering teams that must be consciously guarded against.
• Measure marketing wisely by measuring what you can and valuing what you can’t (investing in brand even without clean metrics, using indirect indicators), anchoring on real economic outcomes (CAC/LTV, customers, revenue, retention) over vanity metrics, resisting measurability bias, and using measurement to iteratively improve — good judgment informed by measurement, not measurement-only decisions.

Further reading

• Purchase funnel — the funnel measurement operates on
• Customer acquisition cost — the key marketing viability metric
• Growth and growth loops (previous post)