Reselling Social Media Services: Seven Checks to Run on a Panel Before Your Clients Rely on It

Many digital agencies that sell social media growth do not deliver it themselves. They place the orders on an SMM panel, a wholesale marketplace for followers, likes and views, and put their own price and their own name on the result. The panel stays invisible to the client until something goes wrong: an order that never starts, a count that drops, a charge for a typo. At that point the agency is the one answering for it. This article sets out seven checks an agency can run on a panel from its public pages before signing up, and what we found when we ran them. It is written by Boostero, an SMM panel, so we are one of the panels measured. Our own results appear in the same form as everyone else’s, including the lines where another panel does better, and every measurement carries the date it was taken.

Most panels run the same software

On 11 September 2026 we read the public API documentation of ten panels, our own included, plus the software vendor’s demo. Nine of those eleven documents could be read; two returned HTTP 403. Eight of the nine served a byte-identical response example, down to the same sample values, because at least seven of the eight panels we could read run on the same panel software, Perfect Panel. So for most panels we read, the ordering system underneath is the same product. What differs is what each panel builds on top of it: its catalogue, its terms, its documentation and its support. That is where the seven checks look.

1. Can you work out your margin before you register?

Your margin is your client price minus the panel’s price, and you can only calculate it if the panel shows its prices to a visitor who is not logged in. Check the services page before creating an account. If you want your own branded storefront, called a child panel, check whether its monthly price is published too. On Boostero both are public: the services list shows each price per 1,000 before registration, and on 13 October 2026 a child panel cost $30 a month. On 17 September 2026 we requested the child panel pages of two other panels at the address a reader would guess, and both redirected to the homepage, so their price sits behind a login. Two of two is not a claim about the market, but any price you cannot see is worth getting in writing before you sign.

2. Is the API documentation the template, or was it measured?

If you plan to connect a panel to your own client dashboard or ordering system, the API page is where you learn how it behaves. Because most panels we read publish the vendor’s template, one quick test separates them: does the page state a dated rate measurement, an error table built from real responses, and the order in which a request is validated? A page that only lists the calls is the template, and the panel has not published what its own instance does. Ours states all three; the rate and the validation order were measured on 4 September 2026.

3. How does order status reach you?

Perfect Panel has no webhooks: the panel does not push status changes to your system. None of the nine readable API documents mentioned one on 11 September 2026, and the vendor’s own feature list names none. Plan for polling from the start. On our panel the status call accepts up to 100 order IDs per request, so one request covers a hundred open orders. On 4 September 2026 we sent 170 sequential requests at about nine per second and every one was accepted; 25 simultaneous requests also succeeded, but the slowest took 2.5 seconds. Polling sequentially at about one request per second, with order IDs grouped, stays well inside what was measured.

4. What happens when a link is wrong?

This is the check that costs agencies money, because the link is where client data enters the order. On our panel, which runs Perfect Panel, a request is checked for the service, then the quantity range, then the balance. The link is not checked at all. On 4 September 2026 we sent a test order with the text “notaurl” as its link. It was accepted, came back with an order ID rather than an error, and was charged. The supplier later reported it completed, and nothing came back to the balance; we read that outcome from the panel’s own record on 13 October 2026. A link that is real but wrong, such as a handle typed with one wrong letter that belongs to another account, is delivered to that link, and on Boostero an order delivered to the link entered is not refunded. Panels differ on this. On 13 October 2026 one panel’s API page said that an order with an invalid link, such as a YouTube link on a TikTok service, can be cancelled and refunded; we did not test that. Read each panel’s own wording, and validate every client link on your side before the order is placed.

5. Read refill as a period, then check it with a count

A refill replaces followers, likes or views that drop after delivery, on services that offer one, for the period the panel states. The place to read it is the service name or description, before you order. We tested it. On 9 September 2026 we ordered 1,000 Instagram followers on each of seven panels, one account per panel, asked each for a refill on 21 September, and counted every account again on 7 October. Six of the seven service names stated a refill term. It was one order per panel, on one platform, over one month, so another service or another month could go differently, but three things came out of it: “Completed” is the panel’s status, not your client’s count. Three panels marked the refill Completed. On two of those three accounts the count on 7 October was still below the count on the day of the request, ours among them: 1,089 on 21 September and 1,021 on 7 October, of which 56 followers were on the account before the order. Acknowledged is not delivered. Six panels acknowledged the request within 1.6 to 7.1 hours. The two largest rises in count came from refills that arrived more than seven days after the request. A longer period did not mean a higher count. The two services sold with 365 days of refill had the lowest and the third-lowest counts on 7 October; the same services were already among the lowest on 21 September, before any request. For an agency the working rule follows: screenshot the client’s count before you ask, ask once with the order ID, and count again a week later before you report back.

6. Does the front page promise what the terms take back?

No panel controls what a platform does to an account, and a careful panel’s terms say so. Some front pages say otherwise. On 13 October 2026 one panel’s homepage said that with its real-quality services “your account is safe and protected from any restrictions”, while its terms said it is not responsible for bans, suspensions or other platform penalties. When the two disagree, plan by the terms; your client will hold you to whatever you repeated. Our terms put it this way: “We are not responsible for actions a platform takes against your account or content, and we do not give refunds or compensation for them.” Read the terms first, and promise your clients nothing the terms do not back.

7. Does the catalogue sell services aimed at someone else’s account?

At least one panel we read sells services that act on other people’s accounts or content. On 13 October 2026 its services page listed report bots for 14 platforms and a reviews group that included negative Facebook reviews. Read the full service list by category before you sign, and know what is in the catalogue before your brand sits on top of it. On 13 October 2026 our public list held 1,659 services, and none of them was of that kind.

What to tell your clients

Purchased engagement supplies signals of a certain type; what a platform does with them is the platform’s decision, and no panel or agency can promise an outcome. Tell clients what is being bought, at the quality the service describes, which refill period applies, and that the count on the account, read on a date, is the measure. An agency that reports that way has the dated count to show when a panel’s status page and a client’s screen disagree.

Where we fall short

We have no live chat: support runs through tickets, Telegram and email at any hour, but there is no chat window to use while an order runs. And in the refill test above, our order page showed Completed while the count on 7 October was lower than on the day we asked. A panel that passes these seven checks has told you in writing what you would otherwise learn from a client complaint. Run them on any panel you consider, ours included.