Why ROAS Lies on a 90-Day Sales Cycle
Want my team to build this for you? Work with Advergent. Rather build it yourself? Join the coaching waitlist.
ROAS lies on a 90-day sales cycle because you read the number before the cash arrives. The click happens today, the booked call a week later, the signed deal months after that. This month's ROAS divides this month's spend by revenue that hasn't landed yet, so it reads far worse than the funnel actually is, and people kill working campaigns because of it.
There are really two separate lies hiding inside one number. The first is timing: you measure a period that's shorter than the time it takes a buyer to pay you. The second is scope: you count the first sale and ignore everything that client is worth afterwards. Both make a profitable account look like a losing one. Here's how to tell what your ROAS is actually saying.
Why does ROAS lie on a 90-day sales cycle?
ROAS lies on a long cycle because the spend and the revenue sit in different months. A high-ticket buyer coming off organic content typically takes 90 to 120 days to buy: they consume content for three to four months before they book. If you divide this month's ad spend by this month's closed revenue, you're comparing today's cost against deals that started before you turned the ads on.
The number that comes out is meaningless in both directions. Early on it reads too low, because you've spent the money but the cash from it is still weeks away, sitting in booked calls that haven't closed. People see a ROAS below 1 in week two and shut the campaign off, when the deals that would have carried it well past breakeven were still three weeks from signing. Later, if you cut spend, ROAS can read artificially high for a month because you're collecting on deals you paid to source earlier.
This is the entire reason paid funnels exist, by the way. A good funnel compresses that 90 to 120 day cycle into 72 hours, sometimes 7 to 10 days, by doing the trust-building that organic does slowly. The faster the cash lands after the spend, the sooner your ROAS tells the truth. If you're fighting a long cycle, read how to shorten a B2B sales cycle before you touch the bidding.
What's the difference between the ROAS Ads Manager reports and your real ROAS?
The ROAS in Ads Manager is modelled and attributed at the click, not at the sale. For a high-ticket offer the purchase happens on a call days or weeks after the ad was seen, so the platform either guesses at a conversion value it never sees or counts a lead event that isn't revenue. Your real ROAS lives in your CRM and your bank account.
For any offer sold on a sales call, I trust the revenue my client actually collected over the period, divided by the spend that sourced those calls, and nothing the platform reports on the revenue side. Ads Manager is useful for what it can see: cost per booked call, cost per click, which creative is winning. It cannot see a $7,000 deal that closed over the phone eleven days after the click, so it shouldn't be the thing you judge the money on.
This is also why cost per lead is such a trap on a long cycle. The platform will happily optimise toward the cheapest form fill and show you a great cost per lead while the real revenue goes nowhere. If that sounds familiar, the fix is in why your Meta ads leads look bad: buy the booked call, not the lead, and measure against cash.
Why does first-month ROAS undercount a subscription or high-ticket offer?
First-month ROAS undercounts because it counts one payment from a client who's worth many. On a subscription that revenue repeats every month they stay, and on a high-ticket offer the first deal is usually the smallest thing that client will ever buy from you. Judge either one on the first transaction and you'll conclude the ads don't work when they do.
Here's the SaaS version, the example I use in my own SaaS video. Say you sell a $79 a month plan and it costs $300 to acquire an account. On day one you're at negative $140 before you even count running costs, negative $210 after, and your break-even is about 3.1 months out. On first-month ROAS, those ads look like a disaster. They're not a disaster, they're a timing problem: the business just needs to survive long enough to collect.
The high-ticket SaaS account behind one of my case studies shows the same shape with real money. It spent $35,627 and that added roughly $95,000 in monthly recurring revenue: about 2.7x in the first month alone, on revenue that then repeats every month those customers stay. A first-month snapshot would call 2.7x "fine." The truth is much better than fine, because month two is nearly free. The whole SaaS economics argument turns on this: you cannot read a recurring business on a single month.
How does your back end change the ROAS you can accept?
Your back end sets how low a front-end ROAS you can tolerate, because your real limit on acquisition cost is dictated by the total revenue a client generates, not the first deal. If you only ever sell someone once, the ads have to pay for themselves on that one transaction. If you have a value ladder, the first sale only needs to not lose money, and the back end does the rest.
I worked with an education business that went from 59k to over 200k a month year over year once they put Meta ads behind a webinar funnel that already closed organically. Their ad budget runs around 45 to 50k a month at roughly 2.5x ROAS. On the front end alone, 2.5x is nothing to brag about. It's a great business anyway, because there's a value ladder behind it that ascends existing clients with no ad spend attached, so the front-end ROAS was never meant to carry the whole thing.
That's the mistake in reading one ROAS number as a verdict. A front-end funnel at 2.5x with a strong back end beats a front-end funnel at 5x with nothing behind it. If your ideal client is worth at least $15,000 in lifetime value, you can pay far more to acquire the first deal and still run a healthy, scalable account. The better the offer and the back end, the more aggressive you're allowed to be on the front, which is the same logic underneath the $1M offer formula.
| What you read | When it misleads | What's actually true | How to read it right |
|---|---|---|---|
| This month's ROAS | Cycle is longer than the month | Today's spend pays off next quarter | Match the window to the cycle, or compress the cycle |
| Ads Manager ROAS | Sale happens on a call, not a click | Platform never sees the real revenue | Use cash collected in the CRM, not reported conversion value |
| First-month ROAS | Subscription or value ladder behind the offer | The client is worth many payments | Judge on lifetime value, not the first deal |
| Front-end ROAS only | Strong back end exists | Back-end revenue sets your CAC tolerance | Count ascension and renewals in the return |
How do you measure ROAS correctly on a long sales cycle?
Measure by cohort, not by calendar. Group the spend from one period, then follow the revenue that specific spend produces until the cycle is actually complete, using cash collected rather than anything the ad platform reports. A ROAS is only finished when the deals it sourced have had time to close, which on a long cycle means the window has to be at least as long as the cycle.
Three rules keep the number honest. First, give a new account real time and spend before you judge it: at least $1,400 over 14 days, because a handful of calls is too small a sample and early ROAS swings wildly on tiny numbers. Second, track cost per booked call as your early signal, since it tells you within days whether the top of the funnel works while ROAS is still catching up to the cash. Third, once a cohort has had its full cycle to close, that final ROAS is the one you trust. In the agency account behind one of my case studies, $35,076 in spend eventually returned $168,000, a 4.79x, but only a window long enough to let those 129 booked calls close would ever show it.
The faster fix is to stop fighting the cycle. The reason I build a 72-hour funnel is partly so the money comes back quickly enough that ROAS stops lagging reality. If you want a system that compresses the cycle and measures the right number from the start, that's exactly what we set up inside coaching and done-for-you, beginning with an audit of what each stage of your funnel is really worth.
FAQ
What's a good ROAS for a high-ticket offer?
Measured over a full sales cycle on cash collected, the accounts behind my case studies ran from about 2.5x to 5.6x and all were profitable. But the number only means something once the window is long enough for the deals to close and once you've counted back-end revenue, not the first sale alone.
How long should I wait before trusting my ROAS?
At least as long as your sales cycle, plus the $1,400 and 14 days of spend any new account needs before the sample means anything. On a 90-day cycle, a 30-day ROAS is still incomplete, because most of the cash it's owed hasn't been collected yet. Use cost per booked call as the early read instead.
Should I use the ROAS in Ads Manager?
Not for judging the money. Ads Manager attributes at the click and can't see a sale that happens on a call days later, so its revenue figures are modelled. Use it for cost per booked call and creative decisions, and use your CRM's cash collected for ROAS.
Why does my ROAS look worse than my competitor's?
Often because you're measuring a shorter window or counting only the first sale. A competitor reporting a higher front-end ROAS may simply have a longer measurement period or a bigger back end doing the work. Compare like for like: same window length, same definition of revenue.