Dunning is one of those words that lets everyone in the meeting nod while picturing something different. Finance pictures overdue invoices. The product team pictures a cancellation flow. Whoever owns billing pictures a queue of failed renewals and a blizzard of automated emails. Vendors, understandably, picture whatever their product happens to do, and the word stretches to cover it.
Before comparing dunning solutions, it is worth naming the work, precisely, because the category only makes sense once the jobs are separated.
The four jobs hiding inside the word
Job one is prevention: everything that keeps a good payment method from going stale in the first place. Card expiry warnings, update prompts inside the product, account updater enrollment working silently in the background. Prevention is unglamorous and, hour for hour, the highest-yield work in the whole stack, because a failure that never occurs needs no recovery at all.
Job two is the attempt strategy: what happens in the first seconds and days after a renewal fails. Which declines get retried, when, how many times, with what spacing, and with updated credentials checked in what order. This is machinery, and it should be judged as machinery: on its logic, its evidence, and its restraint.
Job three is communication: the emails, messages, and in-product prompts that ask the subscriber to act when the machinery cannot fix it alone. Tone lives here, and so does most of the customer-experience risk. A message that arrives before the system has made its own serious attempt reads as outsourcing the problem to the customer, and subscribers can tell.
Job four is the exit: how access is handled during the grace window, when a subscription is finally cancelled for non-payment, and what record the customer is left holding. Done well, the exit preserves the relationship and the winback. Done badly, it manufactures the angriest reviews a billing process can produce.
Why the naming matters at purchase time
Products differ in which jobs they actually perform. Some are communication tools with a retry attached. Some are attempt engines with a template library. Some, increasingly, bundle a cancellation flow and call the bundle retention. None of that is visible in the word dunning, and all of it is visible the moment you score a candidate job by job. The category survey at this dunning solutions comparison organized by coverage is built on exactly that separation, and it is a workable template even if you buy from someone else: list the four jobs, mark what each candidate genuinely does, and price only the coverage you are actually getting.
The naming also protects the jobs nobody sells you. Prevention habits inside your own product, the card-update prompt after a successful renewal, the expiry warning timed to the billing cycle, often sit unowned because the vendor's dashboard starts at the failure. Naming the work assigns it an owner, even when the owner is a thirty-line email your own system sends.
A closing discipline
Write the four jobs on one page, with your current owner and current evidence for each, before the next vendor call. The page will be uncomfortable in at least one row. That row, not the demo, is your buying agenda. Businesses that shop this way end up with smaller stacks and better numbers, because every tool in the stack can answer a simple question: which job are you for, and what did you return last month? For a flat-fee example of a service that answers for the attempt and communication jobs together, Swipe Again publishes its model openly, which makes the comparison easier to run honestly.
Top comments (0)