What Observability Taught Us About Signal Control
This article extends the Deep Dark Funnel series. Start with The Deals You Never Saw if you're new here.
A few years ago, the observability world faced a problem that should sound familiar.
Enterprises were drowning in machine data. Logs, metrics, traces — flooding in from every server, container, and application. The tools they bought to make sense of it all were expensive, siloed, and increasingly overwhelmed. And the data itself was getting noisier every day.
Sound like your GTM stack?
The Observability Problem
The default approach was vertical integration. Buy Splunk. Buy Datadog. Buy whatever platform promised to ingest everything, analyze everything, and surface insights. Hope it works.
But vertical integration had problems.
It was expensive. Splunk's per-GB pricing meant that as data volumes grew, costs spiraled. Companies were paying millions just to store logs they might never query.
It created lock-in. Every tool wanted to be the system of record. Moving to a different platform meant migrating years of data and retraining teams. Nobody switched unless they had to.
And it didn't solve the noise problem. More data came in. More storage got consumed. More dashboards got built. But the signal-to-noise ratio kept getting worse. Teams couldn't find what they needed in the flood.
What Changed
Around 2017, a group of former Splunk engineers saw the structural problem clearly: the bottleneck wasn't in the analysis layer. It was upstream.
Before the data ever reached Splunk or Datadog or Elastic, there was a gap. Raw machine data was flowing in, unfiltered, unprocessed, uncontrolled. Each tool received whatever hit its ingestion endpoint. Nobody governed what went where.
Their insight was simple: if you could control the data before it reached the tools, everything downstream would work better.
So they built a layer. An “observability pipeline.”
It sat between the data sources and the analysis tools. It could filter — dropping the noise before it consumed expensive storage. It could enrich — adding context that made data more useful. It could route — sending different data to different destinations based on actual need.
The key was that it didn't replace any existing tool. It made every existing tool work better.
What the Market Learned
One company that built this — Cribl — grew to billions in valuation. Not because they had the strongest analytics, but because they identified a layer that was missing. Other companies emerged with different approaches to the same gap.
The observability market, over time, learned several things:
You don't have to rip and replace. The most practical solution isn't a new platform — it's a control layer that sits upstream of your existing stack. Keep Splunk for what it's good at. Keep Datadog for what it's good at. Just control what reaches them.
Upstream governance changes everything downstream. When you can filter and enrich at the source, every tool gets cleaner data. Costs go down. Accuracy goes up. Teams can actually trust what they see.
The bottleneck is rarely where you think it is. Companies spent years blaming their analytics platforms for poor signal quality. The problem was never in the tools — it was in the ungoverned data flowing into them.
Control is different from analysis. The platforms that do analysis well aren't necessarily equipped to do governance well. They're optimized for different problems. Trying to make one tool do both creates compromise in both directions.
The GTM Parallel
If you've been reading this series, the parallel should be obvious.
GTM teams are drowning in activity data. Website visits, email opens, ad impressions, content downloads — flooding into the stack from every touchpoint. The tools they bought to make sense of it are expensive, siloed, and increasingly overwhelmed.
And the data is mostly noise. Machines. Bots. Security scanners. Activity that looks like interest but means nothing.
The default response is the same one observability teams tried: buy another tool. Add an intent platform. Stack a CDP on top. Hope vertical integration solves it.
But it doesn't. Because the problem isn't in the tools.
The problem is upstream — in the ungoverned signal flowing into everything.
The Missing Layer
What observability learned is that control and analysis are different problems requiring different solutions. The control layer sits upstream, filtering and enriching and routing. The analysis layer sits downstream, making sense of what it receives.
GTM has built hundreds of analysis tools. Intent platforms. Scoring models. ABM systems. Engagement dashboards.
What it hasn't built is the control layer.
There's no place where raw buyer activity gets filtered before tools interpret it. No place where bots get separated from buyers. No place where context gets added before the signal reaches six different tools that will interpret it six different ways.
That gap — the same gap that observability addressed a decade ago — is where the deep dark funnel lives.
Why This Matters Now
The observability companies didn't address the problem by building better analytics. They addressed it by identifying a gap and building there.
Whether GTM follows the same path is an open question. The parallel is structural, not guaranteed. Markets don't always learn lessons from adjacent spaces, even when they should.
But the diagnostic is worth considering: if the problem is upstream, no amount of downstream optimization will fix it. That was true for observability. It may be true for GTM.
What form the answer takes — if it takes any form at all — remains unclear. Some companies may address it through operational discipline. Some may build internal capabilities. Some may wait for new categories to emerge. Some may decide the problem isn't worth addressing.
The point isn't to prescribe an answer. It's to notice that the question is being asked — and that a similar question, in a similar domain, eventually found a structural response.
Next in the series: What Context Actually Means
Subscribe for new articles as they publish.