That new CARR smell

With politics and the pandemic hurtling towards a crescendo this week, I thought it would be a welcome distraction to talk about something equally as controversial… Contracted Annual Recurring Revenue (CARR). (The comparison is mostly a dark joke, so spare me any indignation.)

We started seeing CARR enter the startup lingo about five years ago, but typically only inside of enterprise software startups in their tracking of implemented revenue versus signed revenue. Then CARR started appearing in investor decks, then companies wanted to be valued as a multiple of it instead of ARR, then companies started talking about “verbal” Contracted Annual Recurring Revenue (VCARR). Oh boy.

Here is how we think about ARR and CARR and how they relate to each other:

Your current GAAP (recognized) monthly revenue x 12, excluding any transactional revenue

= ARR (very conservative)

+ transactional revenue that is highly certain to recur (example would be “metered” subscription models like say charging $1 per truck load in shipping)

= ARR (nominal)

+ additions to ARR expected in the next 6-12 months (or maybe before end of calendar year) that represent increases baked into contracts that are already implemented. This could include highly likely increases to recurring transactional revenue for certain models

= ARR (liberal)

+ “ARR” that is contracted but not yet implemented or billing, representing what will be billing immediately after it is implemented

= CARR (nominal)

+ increases to CARR that represent contracted annual price increases over time in contracts not yet implemented, including future “recurring transactional” revenue

= CARR (liberal)

+ contracts you think you have a verbal on

= VCARR (bullshit) = bullshit

We generally prefer the use of ARR (nominal) and CARR (nominal) both in working with our companies and in evaluating new investment opportunities. We prefer these versions because their constituent parts are more transparent and objective. The difference between them is also very clear. ARR is monthly revenue today x 12, and CARR is monthly revenue if we finished all implementations today and multiplied by 12. We loathe VCARR.

But don’t you want to take credit with your BoD or an acquirer/funder for everything you have signed? Of course, but if there are too many assumptions in the underlying structure of a metric, its veracity collapses, potentially requiring backtracking, reforecasting or undermining your story with a BoD or acquirer/funder. Using nominal definitions, it is okay to say our ARR is X today, but in six months, we already have Y baked in, representing contracted increases. Similarly, many companies will talk about TCV, which includes the value of multi-year contracts.

These days, we are seeing most companies valued by some multiple of ARR (nominal) or expected year end ARR (nominal). If the former, credit is usually given also for unimplemented contracts (CARR – ARR). If the company is later stage with a long history of successfully implementing contracts, there may be no discount on the multiplier used for the unimplemented contracts. For an early stage company say with $1M in ARR and $1M in unimplemented contracts (CARR = $2M), the discount on the unimplemented multiplier might be as high as 50%, given more uncertainty about implementation success.

What matters the most is that whatever metric you choose, it is clearly defined, transparent and able to be objectively measured. Otherwise your new CARR will smell fishy.