Picking the right metrics for a tech business is a lot harder than people think
Most technology companies I see stumble over their KPIs because they start with vanity numbers. Revenue looks good on a dashboard. Total users sounds impressive in a pitch. Neither of those tells you whether the business is actually healthy or just spending aggressively to grow. The problem compounds when leadership picks six different metrics and expects the whole company to align. It never works. I learned this the hard way about four years ago when I was brought in to audit a Series B SaaS company. They had seven dashboards across five different tools and everyone was looking at something different. The CEO was watching top-line revenue. The head of product was obsessed with daily active users. The finance team had their own model running on quarterly ARR with completely different assumptions about what counted as a "contract." We spent two weeks reconciling the numbers and found they were off by 18 percent on net revenue retention because of how each department handled credits, churn, and expansion revenue. That 18 percent gap is the kind of thing that quietly destroys credibility with investors when it surfaces during a due diligence call. The fix wasn't adding more metrics. It was picking the right ones and locking down exactly how each was defined, who owned it, and how often it was refreshed. I wrote up a single page that listed the five metrics every tech company I've worked with should track and how each one should be calculated. No ambiguity. "Gross revenue minus refunds and credits, reported monthly" instead of vague language that lets each department interpret it their own way. This typically cuts dashboard reconciliation time from about two weeks down to a couple of hours and eliminates the kind of discrepancies that make investors uncomfortable.
The core set is relatively small. Net dollar retention comes first because it captures everything—churn, downgrades, upgrades, and expansion—in a single number. A company with 120 percent NDR is growing efficiently even if total revenue growth looks modest. Below 100 percent means you're losing more to churn and contraction than you're gaining from existing customers. That's a structural problem no amount of new customer acquisition can fix. Rule 40 score is another one people misunderstand. It's not MRR growth rate plus gross margin. The traditional definition is revenue growth rate plus free cash flow margin. You add those two percentages together and if the result is 40 or higher, the company is generally considered healthy by SaaS benchmarks. It's a quick health check, not a sophisticated valuation tool. But it's fast, it requires no complex calculations, and it catches companies that look fine on growth alone but are bleeding cash to get there. CAC payback period deserves more attention than it gets. This measures how many months it takes for the gross profit from a new customer to cover the cost of acquiring them. Under 12 months is strong. Over 24 months is a warning sign that your growth model is capital inefficient. I once worked with a company that had 31-month CAC payback and still kept scaling sales spend because their revenue growth looked great. They ran out of runway 14 months later. The metric would have flagged this immediately.
Lifetime value to CAC ratio is the companion piece. You want this to be at least 3 to 1. Below that, you're spending too much to acquire customers relative to what they generate over their relationship with you. Above 5 to 1, you might be underinvesting in growth. The math is straightforward but the inputs matter enormously. LTV depends heavily on how you estimate customer lifespan and average revenue per user, and small changes in those assumptions can swing the ratio dramatically. Churn rate needs to be separated into logo churn and revenue churn. Logo churn tells you how many customers leave. Revenue churn tells you how much revenue leaves. A company can have low logo churn but high revenue churn if enterprise clients are downgrading or walking away. Those are completely different problems with different solutions. Fixing logo churn usually means improving onboarding or product fit. Fixing revenue churn often means addressing pricing, feature gaps, or competitive pressure on larger accounts. There's also a practical implementation side to all of this that most guides skip. You need a single source of truth for revenue data. If your CRM says one thing and your billing platform says another, your metrics are lying to you. I recommend linking your CRM to your billing system through a middleware layer rather than syncing directly. This catches discrepancies in real time instead of discovering them during quarterly close. Setting this up usually takes about three weeks of engineering work and saves roughly ten hours per month in reconciliation time going forward.
Get the Full Details
Data freshness matters more than sophistication. A simple metric updated daily is more useful than a complex one refreshed monthly. Your team should be able to look at a dashboard in the morning and act on it the same day. If the data is a month old, it's historical reporting, not operational guidance. Monthly reporting is fine for investor updates. Daily or weekly is what drives decisions. One specific thing to watch for is cohort-based analysis. Aggregate numbers hide a lot of problems. A company might show 10 percent MRR growth while the most recent quarter's cohort has 25 percent churn. Looking at the overall number would miss that entirely. Break down your key metrics by cohort—by sign-up month, by acquisition channel, by plan tier. The patterns that emerge usually tell you exactly where to focus your energy. Another counter-intuitive point: fewer metrics are better. I've seen teams track over 30 KPIs and barely glance at more than five of them each week. The rest become noise. Pick the five or six that directly drive your biggest decisions and build the rest of the organization around those. Everything else is secondary. When people resist this, it's usually because they're afraid of missing something important, but the research on decision-making bandwidth supports the narrower approach.
Here's where the limitations come in. These metrics don't capture everything. They won't tell you about employee morale, product quality issues before they hit retention, or competitive threats that haven't shown up in the data yet. NDR is backward-looking. It reflects what already happened, not what might happen next quarter. You need qualitative signals alongside the quantitative ones—customer interviews, support ticket trends, product usage heatmaps—to fill the gaps. None of these KPIs are a complete picture on their own. The other honest limitation is that benchmark comparisons are only useful within your segment. A 120 percent NDR looks great for an enterprise SaaS company but might indicate underperformance for a low-touch consumer app. Market cap, business model, pricing strategy, and growth stage all change what "good" actually means. The Rule 40 benchmark itself comes from Benchmark's SaaS research and applies broadly to venture-backed software companies, not every technology business out there. Adapting these to your specific context is necessary, not optional. If you're starting from scratch and don't have the engineering resources to set up proper data pipelines, consider using a tool like MetricWire or Dryrun specifically for SaaS metrics rather than trying to build everything in-house. They handle the standard calculations correctly and reduce implementation time to about two weeks instead of two months. The trade-off is a monthly cost that scales with your revenue, usually between $500 and $2,000 depending on company size. For most early-stage companies, that cost is worth avoiding the reconciliation mess that happens when finance and sales define metrics differently.
The single most important thing is to document every definition in writing. Not in a slide deck. Not in a conversational Slack message. In a shared document that gets updated when definitions change. Version control the document. When you revisit these metrics six months later—and you will—you need to be able to see exactly how the definitions evolved. Without that record, comparing periods becomes guesswork, and the whole exercise loses its value.
