What to settle before you buy MCA leads

Start with the seats, not the catalogue

The first number in this transaction is not a price and it is not a record count. It is how many records your floor can physically touch before the campaign is over.

Work it out with your own figures. Take the seats that will genuinely have this file loaded — not your headcount, the seats — times the dials one of them gets through in an hour on your dialer, times the hours a day they spend on it, times the days you are giving the campaign. Whatever that comes to is the ceiling. Everything above the ceiling is inventory you paid for and then aged yourself.

That is the real cost of overbuying and it never shows up on the invoice. Surplus rows do not wait politely at the bottom of the queue. They get worked in week three by a rep who has already learned that the good ones went first, and then the file earns a reputation on your floor that the file did nothing to deserve.

The same arithmetic tells you which product to ask about before you have asked about any of them. Two closers and a long cycle is a depth problem. Twelve seats judged on conversations is a breadth problem. Any vendor will sell you either one. Only one of them fits the room.

Make the first order readable

A first order is an experiment, and an experiment with four things moving inside it produces a result you cannot attribute to anything and an argument nobody wins.

So move one. One product, one set of filters, one window, the script you already run. Change the states, tighten the revenue band and hand the batch to your two best closers all at once, and what you have learned at the end of it is that your two best closers are good. You knew that.

Settle the dispositions before the file lands, in words rather than in codes nobody remembers by Thursday. What the list has to do is keep the line did not work apart from the merchant did not bite, because those two facts go to different people afterwards: one comes back to the desk you bought from, the other stays on your floor. A column that only ever says no cannot be split open later.

Then size it so the answer means something. A batch small enough to be free is usually too small to read — a couple of bad rows in a thin cut will swing the percentage far enough to confirm whatever you already suspected. Size the test to roughly a week of one seat, run it without touching anything, and read it once at the end.

Open the file before anyone dials it

The cheapest twenty minutes in the campaign are the ones between the file arriving and the first call going out. In a spreadsheet, before the floor ever sees it:

• Count the rows against the count you approved. They should match. A file that came back longer than the count is a question, not a bonus.
• Sort on the phone column, then on the business name, and look for the same merchant twice. Duplicates inside one order are a defect and they are easiest to raise on day one, while nobody has dialled anything.
• Run it against your own system. What you already own, already worked or already suppressed comes out here, and the number left standing is the number you actually bought.
• Check the columns landed where your importer thinks they did. A state that imported into the city field will survive all the way to a confused merchant.
• Dial ten rows yourself. Not a rep — you. Ten calls tell you more about a build than the specification did.

Do it on every delivery, including the fourth one from a desk you have come to trust. It costs twenty minutes and it turns a vague feeling in week two into a claim in hour one, which is the difference between a replacement and a grudge.

Week one answers two different questions

Two things come off the first week of dialling, and keeping them apart is most of the skill in buying data.

The first is whether the row reaches the business printed on it. A dead line, a stranger with no connection to the name, a business that shut for good — those are the file’s failures. Count them precisely, because they are the only ones anybody owes you anything for.

The second is what happens after a merchant says hello. No answer, not interested, took money somewhere else, does not remember filling anything in: none of that is a defect. It is the job. A file that reaches the right businesses and still converts badly is telling you about your offer, your calling hours or the state of that merchant’s last quarter, and none of those things live in a spreadsheet.

Blur the two and you lose twice. The first order goes on arguing about the wrong number, and you never build the one asset that makes the second order better: a record of which filters put your closers in front of businesses worth talking to.

What aged and exclusive are really selling

Both words describe the same quantity from opposite ends — how much of this merchant’s attention somebody else has already spent.

Exclusive controls the half of that a vendor can control. One buyer per record, out of inventory the day it ships, so it cannot come back around next quarter wearing a different product name. That is a mechanic, and you are entitled to ask how it is enforced. What it cannot do is tell you the merchant has not filled in four other forms this month, because nobody knows that. A vendor who implies otherwise is charging exclusive money for a guess.

Aged runs the other way. The attention has already been spent and the price says so. It earns its place when the seats are being paid for whether or not the dialer has anything queued, and it stops earning it the moment somebody expects fresh behaviour at aged prices. That is a question about your cost per hour, not about the data.

Between the two sits the variable you own outright and have probably never measured: how long a record waits between landing in your system and its first dial. Pull that number this month. It moves your results more than the tier on the invoice does, and unlike the tier, it costs nothing to change.

The language on the file becomes your problem

Some vendor copy should stop an order rather than close it. A file sold to you as “pre-approved”, “pre-qualified” or “credit-verified” is not a better file — it is a contact file being described as a credit decision. The same goes for “guaranteed approval”, and for “underwritten leads”. The rows underneath may be identical to the rows in the file next to it. The words are what moved the product into another category.

That category has a regulator in it. The Federal Trade Commission’s long-running case against Trans Union over its target-marketing lists is the one every lawyer in this market reaches for: lists assembled and sold for marketing were held to be consumer reports, with all of the obligations that follow. The data had not changed. What it was sold as had.

And the exposure is not only the seller’s. You are the one who puts those records in front of a funder, and the one holding a file that was marketed as a determination about credit. A contact record says which business went looking for money and how to reach it. It says nothing about what that business can actually get — that is what your underwriting is for — and no adjective on a vendor’s website changes what is in the row.

Treat claimed compliance the same way. A registry check happened on a morning and describes the file as it stood that morning; it is not a quality baked into the rows, so ask when it ran and which products it ran on. Past that point, what you send, to whom, at which hour and on what dialer setting is your call, your exposure, and a conversation for your own counsel rather than for anybody’s marketing page.

The short version

1. Find the ceiling first. Seats, times dials an hour, times hours a day, times days. Buy under it.
2. Move one variable. Product, filters or script on the first order — not all three at once.
3. Fix the disposition words before delivery, so a bad line and a bad call never share a column.
4. Spend twenty minutes in the spreadsheet before anyone dials: count, de-dupe, map, spot-dial.
5. Split week one into what the file did and what the pitch did, and claim only the first.
6. Measure the wait before first dial. That number belongs to you and it is free to fix.

None of that needs a vendor’s cooperation, which is rather the point of putting it here. The parts that do are already on this site: the sample file carries the full column list with nothing standing in front of it, and the replacement policy prints which defects get swapped and which do not, in the words we use on the phone. Read both before you talk to anybody in this market, ourselves included.