What was the first sign, looking back, that the network was going under

Lost about $6,200 in May to a network that had paid me on time for fourteen months straight, and the thing that bothers me isn't the money. It's that I've gone back through six months of emails and the signs were all there and I read every one of them as normal.

AM changed in January. Two of my best offers 'paused pending advertiser review' in February and never came back. March payment came four days late with an apology. April came eight days late with no apology. May never came, then the site started 502ing, then the telegram went read-only.

Every one of those has an innocent explanation and I gave it one at the time.

So what I want from this thread is the specific things people noticed before it went. Not 'trust your gut'. The concrete tells - what changed on the platform, in the emails, in the offer list, in the payment behaviour. I'd like to end up with something I can actually check monthly.

late once is admin. late twice in a row is cashflow. there's your whole checklist

I've been on the wrong end of three of these since 2019 and I keep a list precisely because of what you describe - each one is explainable alone and only obvious in a group.

Caveat: two of my three were small networks. A big one folding looks different and I've never seen it up close.

What I actually watch, in rough order of how early it shows up:

1. The offer list shrinks from the top. Not random offers going - the best-paying ones going first, because those are the advertisers who demand prepayment and the network can no longer prepay. This showed up four months early in two of my three.

2. Payment terms get 'reviewed'. Net15 becomes net30 for 'alignment with industry standard'. Nobody ever moves payment terms in the affiliate's favour, and nobody moves them at all when things are fine.

3. Minimum payout goes up. From $100 to $250 to $500. Every step is money they get to keep on the balance sheet another cycle, and it is always announced as a platform improvement.

4. The AM turns over and the new one has no history with the advertisers. This is the one that gets argued about and I'll defend it below if anyone wants.

5. They start pushing a specific offer very hard. Late-stage networks lean on whichever advertiser still pays them fast, and you can feel it in the emails.

6. Support response time goes up while the tone gets friendlier. The friendliness is the tell. People who can pay you are brisk.

If two of these land in the same quarter I stop building on the network. If three do, I take everything down to the minimum I can and stop reinvesting.

saving this one

Points 1 to 3 are observable. Points 4 and 6 are vibes with a number attached. Which quarter did each of yours actually fire in?

The only one of these I can put numbers behind is payment lateness, and it turns out to be enough on its own.

Two networks that failed on me, days late per cycle, counting from the stated payment date:

Cycle	Network A	Network B
T-6	0	0
T-5	0	1
T-4	2	0
T-3	1	3
T-2	4	5
T-1	9	11
T	never	never

T is the cycle that didn't arrive. Both of them show the same shape: noise, then a step, then a doubling, then nothing. The step happens three cycles out in both cases.

Against that I have four networks I'm still with, and across 71 cycles the worst run is two consecutive cycles at 1 and 2 days, which resolved. So grindstone's rule is close but slightly wrong: it isn't twice in a row, it's twice in a row and getting worse. One day then two days is admin. Four days then nine is a queue with a priority order and you're not at the top of it.

Log the number of days every single cycle. It takes five seconds and it is the only leading indicator in this thread that doesn't require interpretation.

oh no i have one at 3 days then 6 days

Two tells that aren't on Marta_K's list and that I'd rate highly.

First: they stop taking new affiliates, or the opposite - they start approving anyone. Both are bad and they mean different things. Closing the doors means they can't fund more payouts. Opening them wide means they need volume urgently enough to accept traffic they'd have refused six months ago, which is how the advertiser relationship dies next.

Second: check whether the offers in the list are still live at the advertiser. Late-stage networks leave dead offers in the catalogue because pruning them makes the list look thin. Click through to three landing pages a month. If one of them redirects somewhere odd or the advertiser's own signup is gone, that offer has been dead for a while and nobody noticed, which tells you how many people are still working there.

The first sign is that you signed up.

One that happen to me twice my friend and nobody say it yet. The network start to pay in a different way than before, and they say is more convenient for you.

First time they move from bank to a e-wallet that i never hear about. Second time they say now we pay in crypto only, is faster. Both time is not about convenient, is because the bank close the account. Two month later both is gone.

So when the payment method change and you did not ask for it, that is the moment to take everything out.

I've watched maybe nine of these go down since 2012 and there's a pattern underneath all the individual tells that I think is worth naming.

A network is a float. They get paid by the advertiser on net45 or net60, they pay you on net15 or net30, and the gap is funded out of working capital. Everything on Marta_K's list is a symptom of that gap widening. Minimum payout rising, terms extending, offers vanishing from the top - those are all the same event seen from different angles, which is the same event: money is going out faster than it's coming in.

What changed over the years is only how it presents. In 2013 they just stopped answering the phone; there was no telegram, there was no status page, and one Tuesday the ICQ went dark. Around 2016 they started sending the 'we are upgrading our payment infrastructure' email, which I have now received five separate times and which has never once meant anything other than the bank pulled the account. The current version is the one where a 'restructuring' is announced and a new entity name appears on the invoices about six weeks before it folds.

The genuinely useful thing: watch the entity name on your invoice. Not the brand, the legal name at the bottom. When that changes, the old one is being left behind with debts in it, and one of those debts is you.

the invoice entity one is actually good. everything above it is the same paragraph four times

There's an infrastructure side to this that's cheap to check and nobody does it.

Depends what you're running, but for any network you've got real money sitting with, once a quarter go and look at their domain. Three things:

1. Renewal date. A company planning to exist renews in years. A company that isn't renews in one-year increments, and a company in trouble lets it get inside 60 days before renewing. If the expiry keeps rolling forward by exactly twelve months at the last minute, somebody is watching the card.

2. Nameserver or host changes. Moving off a managed platform onto something cheap is a cost cut, and cost cuts at a network hit infrastructure before they hit salaries.

3. Whether the tracking domain and the corporate domain are still on the same setup. When they split and the tracking side goes somewhere cheaper, that's usually the same quarter the ops team shrank.

None of these is proof. All three moving in one quarter, alongside anything on Marta_K's list, and I'd be pulling my balance down weekly.

That one's checkable. Thanks.

Coming back to your challenge on points 4 and 6 because you're half right and I want to be precise about which half.

On timing: network one, AM changed nine months before it folded, which is useless as a signal. Network two, AM changed in the same month the offers started disappearing. Network three, no AM change at all - the same person was there the day it stopped paying, and she was still apologising to me by PM a fortnight afterwards, which I don't think she had to do.

So one of three fired late, one fired alongside a better signal, one didn't fire. You're right that it's weak on its own. What I'd keep is the narrower version: an AM leaving who had been there years and had relationships with the advertisers, replaced by someone who talks about the platform rather than the offers. That's not turnover, that's the institutional knowledge walking out, and it's usually the last thing to go before the money does.

On point 6 I'm not backing down. Support getting friendlier while getting slower is a real pattern and it's because the brisk people have left and the ones remaining are managing you rather than fixing anything. I can't put a number on it. I still stop reinvesting when I feel it.