How to evaluate a nutra CPA network before scaling spend

You don’t judge a network by its landing page or the payout shown on the offer wall. You judge it by six levers: payouts, hold and tracking discrepancies, capacity, GEO coverage, support, and individual terms. Every one of them gets checked before the large launch — on a small test, in written agreements, and against the network’s public track record. Below is a checklist for filtering out a weak network while the cost of a mistake is still a test budget and not a month of margin.

Why you evaluate a network before volume, not during

On a small test, a weak network looks almost identical to a strong one: leads flow, the approval rate looks alive, the manager replies. The difference shows up at scale, where every weakness gets multiplied by volume. A payout delay you could tolerate on a test becomes a cash-flow gap at scale: the next budget is already running while the money for the previous one hasn’t arrived. A discrepancy in credited conversions, invisible on a hundred leads, eats your margin on tens of thousands. A cap you never tripped over suddenly becomes the growth ceiling. Moving volume back out of a weak network always costs more than never sending it there in the first place.

One clarification on scope: this article is about choosing a network for volume. How to grow volume inside a network you’ve already chosen is a separate topic: how to scale nutra campaigns.

Payouts: speed, methods, minimums

What to check. Real payout frequency and speed, methods available for your jurisdiction, minimums, fees and — separately — how the network behaves at month-end and on holidays, when the payout queue is longer than usual. The schedule on the website and the actual time money lands are not always the same thing.

How to check. Request the payout policy in writing before launching and compare it with the network’s public financial policy. On your first small volume, run one full cycle from period close to money received, and measure the actual time against the promise — it’s the only way to see real speed rather than a declaration.

How it works with us: Shakes pays daily. New partners get their first payouts automatically once a week from $200, and early withdrawal unlocks after three payouts. The mechanics are detailed in our work rules and financial policy.

Hold and tracking discrepancies

Two mechanisms that are most often misunderstood, so let’s start with definitions — neutrally, as working mechanics, not as accusations.

Hold is deferred conversion confirmation: the network holds a lead until its quality is verified (in COD — until the order is actually paid on delivery). It’s a normal part of the model, not money being kept “just because.”

Discrepancy — what the industry calls shave: the difference between what your tracker counted and what the network’s dashboard confirmed. Part of it is explained by technology — lost postbacks, deduplication, leads rejected for quality — and part becomes a topic for a conversation with your manager. A discrepancy by itself is not a verdict; it’s a metric you need to measure.

How to check. This is the core of the evaluation, and it’s done by reconciliation, not by trust:

• Reconcile your own tracking stats against the network dashboard on a control volume — by day and by GEO, so you see the stability of the approval rate rather than a lucky or unlucky slice.
• Look at whether the network gives reasons for rejected and held leads or shows only the final number.
• Connect postback statuses so you can reconcile confirmations in real time, not after the fact.
• Evaluate the length and predictability of the hold, not just its existence: a stable, understandable hold is manageable; a floating one without explanations is not.

We deliberately quote no typical hold or discrepancy percentages: they depend on the offer, the GEO and the traffic source, and any “average numbers” would be misleading.

Capacity and caps: can the network handle your volume

What to check. Daily and weekly caps on the offers and GEOs you need, the capacity headroom for growth, and how fast caps are raised against proven volume. A network can cover your current volume and hit its ceiling exactly where you planned to grow.

How to check. Request the caps for your target offers in writing — before launching, not after hitting the wall. During the test, watch whether your flow reaches the ceiling earlier than planned. Separately, ask about the expansion procedure: by which metrics the network raises caps and how fast — in hours, in a day, or “when something frees up.” The answer separates a network ready to grow with you from one with a fixed limit.

GEO coverage for your needs

What to check. Whether the network has strong offers in your GEOs and verticals specifically, not “70 countries in total.” A showcase country count says little: what matters is depth — several competing offers per GEO instead of one token offer, local landing pages, languages, and a call center for COD.

How to check. Match your list of priority GEOs against the network’s offers and ask directly which GEOs are currently its strongest in nutra — network priorities shift. On the test, compare the approval and purchase rates in your GEO against expectations. Separately, check the call-center quality in your target countries: in COD it’s the phone call that largely determines the purchase rate, and a weak local call center kills even good traffic.

How it works with us: an advertiser appears on the Shakes showcase only after screening and a test — with confirmation statistics, a mandatory deposit and a working call center. That’s why our “70+ GEOs” is not a showcase country count but 3,000+ offers from 50+ vetted advertisers, many of which run deep coverage with steady volumes in their working regions. Your manager will tell you the depth for a specific offer–country pair.

Support and a personal manager

What to check. Response speed and quality, whether you get a dedicated manager at your volume, their authority (do they resolve cap, rate and disputed-lead issues themselves or only forward them upward), working timezone and language. At volume, a manager is reaction speed on a problem — and every hour of downtime costs money.

How to check. Before launching, ask two or three substantive questions and measure not just the speed but the quality of the answer — on point or a brush-off. Ask whether a personal manager is included at your volume level and from what threshold. Check whether a disputed case gets driven to a resolution rather than left hanging.

What a manager unlocks at the start and how access levels work — in a separate article: how to start working with a CPA network: access levels and first steps.

Individual terms

What to check. The network’s readiness for an individual rate and rate increases against proven volume, buyback of test spend, contractual terms and early payouts. These are the levers that separate a network “for scale” from a network “for everyone”: base terms are comparable across many networks, but the difference at volume accumulates exactly here.

How to check. Ask directly which terms are available for your volume and from what threshold they switch on. Ask for agreements to be fixed in writing. And evaluate the package as a whole: a rate increase must not be cancelled out by a quietly tightened hold — the levers are assessed together, not separately.

More on individual rates, contractual terms and early payouts for large partners: individual terms for large partners.

The checklist before moving volume

1. Payouts — policy received in writing, one full cycle run on a test, actual timing matched the promise.
2. Hold and stats — your stats reconciled with the dashboard by day and GEO, rejections are transparent, hold is predictable.
3. Capacity — caps for target offers are known, headroom for growth exists, the expansion procedure is clear.
4. GEO — strong offers in your GEOs, coverage depth and call center verified on a test.
5. Support — you have a manager at your volume, they answer on point and drive disputed cases to resolution.
6. Individual terms — thresholds for rate increases and contractual terms are stated and fixed in writing.

No single item is evaluated in a vacuum: weak payouts are not compensated by great caps, and a generous rate increase is meaningless under a floating hold. A working decision is all six levers aligned with your model: GEO, source, vertical and launch tempo. A network that’s strong on paper but misaligned with your needs will lose exactly where you planned to grow.

The same approach applies to choosing the offer itself: not by the promised approval rate, but by verifiable data — how to choose a nutra offer based on real data.

Want to test this checklist against live terms — bring your numbers for a review: we’ll show you the payout policy, caps and terms for your volume at Shakes and check them against the points above on your test.