"...some words of wisdom Mike Arrington once shared with me. He said “numbers always ruin a good story.”

What Mike meant by this is you can raise a seed (or Series A) on a story. But at some point, you will have numbers; users, user growth, revenues, and revenue growth. You will also have a burn rate. And those numbers will become the thing you are judged on and your nice story will be “ruined” by the numbers." --

Key Metrics

(and typical values)

What you want to do is get 'results' and increase the probability/amount of 'success', but you can't quantify or measure 'results' and 'success' directly.

Good characteristics of a KPI (key performance indicator) metric:

In addition, a metric should be something that you'd like to adopt as a goal. Like any goal, it should:

People recommend focusing on 3-5 metrics.

There are some financial metrics (such as revenue and profit) which are important across industries. Beyond these, different industries and business models have different metrics which are most important. It's worth trying to find out what metrics other companies in your industry, or who have your business model, use.

Basic financial metrics

SAAS metrics (software-business-specific)


Consulting metrics

I hear that consulting often uses revenue, rather than profit, as a metric for partners. I'm not positive why this is revenue and not profit, but presumably it's because, as consulting is not a scalable business model, the partner can't do much about costs; the consulting firm is billing by the hour, and the primary cost to the firm is people's time, and a given person assigned to a project will have a set rate that they are paid.


"In SaaS?, you’d try to get a 3X ratio for CAC:LTV" -- [1]

" Key Financial Indicators

All three indicators are important, but emphasis shifts depending on stage:

    Cash (Early Stage: Pre- Product/Market Fit)
    Run Rate (Growth Stage: Post- Product/Market Fit)
    Annual Revenue (getting ready to IPO)"

cash (and runway):

run rate:



" Some of the most common SLAs I have seen used are:

    Availability: the percentage of the time the service is operational. While it is tempting to want to have a system that has 100% availability, achieving this can be really difficult, as well as expensive. Even large and critical systems like the VISA card network, Gmail or internet providers don't have 100% availability - over years, they will be down for seconds, minutes or hours. For many systems, the four nines availability (99.99%, or about 50 minutes downtime per year) is considered high availability. Just getting to this level is quite the work usually, to get to.
    Accuracy: is it ok for some of the data in the system to be inaccurate or lost? If so, what percentage is acceptable? For the payments sytems that I worked on, accuracy needed to be 100%, meaning no data was allowed to be lost.
    Capacity: what expected load should the system be able to support? This is usually expressed in requests per second.
    Latency: in what time should the system respond? What is the time that 95% of the requests and 99% of the requests should be served? Systems usually have a lot of noisy requests, hence the p95 and p99 latencies are more practical usage in the real world.

" -- [2]