Two different questions get asked with almost the same words, and only one of them has an obvious answer.
How do we increase ARR is about the number going up. There are four levers, most companies lean on the worst one, and the fix is mechanical.
How do we improve ARR is a different question, and most companies have never asked it, because ARR arrives in the board deck as a single figure and a single figure has no quality to inspect.
It has plenty. You can increase ARR and make it worse in the same quarter, and almost nobody notices until the year after.
What worse ARR looks like
Same number on the slide. Different thing underneath.
Discount-laden. You hit the number by giving away price. The revenue is real and the margin isn't, and you've taught a cohort of customers what your list price actually means, which is that it's negotiable if they wait.
Concentrated. A meaningful share sits in a handful of accounts. That's not growth, it's exposure, and it reprices your whole company the day one of them leaves.
Acquired at a rising cost. The number went up and what it cost to make it go up went up faster. That works until funding conditions change, and then it stops working immediately.
With no path to a second sale. You sold everything at once. There's nothing left to present when the customer succeeds, so a customer doing well produces exactly the same revenue as one barely hanging on. That's a design choice, made accidentally, that caps the account permanently.
Renewal-dependent. The base holds because switching is painful rather than because staying is valuable. Retention looks fine. What it's measuring is friction, and friction has a shelf life.
Every one of those produces a healthy-looking ARR figure. None of them survives contact with a bad year.
Why it doesn't get looked at
Because ARR is reported as a total, and a total is designed to hide composition.
You'd never accept that anywhere else in the business. Nobody reports "expenses" as one line. Nobody reports headcount without saying where the people are. ARR is uniquely permitted to arrive as a single number that gets compared to last quarter's single number, and the comparison feels like analysis.
Two companies can post identical ARR and identical growth. One of them expanded existing accounts at full price with a clear next step. The other discounted its way into new logos it will spend next year defending. Those are not the same company, and the metric everyone is looking at cannot tell them apart.
The composition test
Four questions. They take an afternoon with a spreadsheet, and most executives have never seen the answers for their own business.
How much of this year's ARR came from customers who were already customers? Not renewal, expansion. New money from people who already paid you. If that number is small, everything you have is coming from the most expensive lever you own.
What did you discount, and to whom? Sort by discount depth rather than by size. If your largest customers carry your deepest discounts, your revenue concentration and your margin concentration are the same problem wearing two hats.
How much revenue sits with customers who have nothing left to buy? That's your ceiling, made visible. Those accounts can only stay flat or leave.
What share of the base is holding because leaving is hard? You know which accounts these are. Everyone does. Nobody writes them down, because writing them down makes them a number somebody has to own.
What improving it actually means
Improving ARR means changing where it comes from, not how much of it there is.
It means the next dollar is more likely to come from an existing customer than a stranger. It means fewer customers who've bought everything you sell and more who have a genuine next step available. It means the base holds because it's working rather than because leaving is expensive. It means price holds because value is visible, not because a negotiation went well.
None of that shows up as a bigger number this quarter. Some of it shows up as a smaller one, because the discount you didn't give is revenue you didn't book, and the deal you didn't force is a logo that isn't on the slide.
That's the part that makes this hard, and it's the reason it rarely gets done. Improving ARR asks somebody to accept a worse-looking quarter for a better-looking business, using a metric that can't display the difference.
Where to start
Run the four questions. Don't fix anything yet.
Most of the value in that exercise is discovering that the number everybody manages to has never been examined, and that the parts of it behave completely differently. Once you can see the composition, the decisions about what to change get much less interesting, because they get obvious.
The number tells you how you did. The composition tells you whether you can do it again.
Lincoln Murphy formally named and popularized Customer Success starting in 2010 and has spent 15 years connecting it to expansion revenue and commercial outcomes. Read The Premise.