Ninja Shield + Lead Rating + Smart Bidding

You pay for clicks that never happened. And Google bids for forms, not for customers.

Those are the two holes in any lead-generation account, and they feed each other: fake traffic reaches the form, and the algorithm learns from it. Ninja Leads is three layers that close them in order — fraud first, quality second, bidding last — so that Google Ads knows what each lead is actually worth and, above all, bids with that data.

⚙️ The engine runs on our server 🏷️ Installed with one tag 🔒 Consent built in
Layer 1 · before the spend

🛡️ Ninja Shield

Watches, scores and blocks traffic that isn't an interested person — with six layers of signals and a farm registry shared across the whole fleet.

  • Placements, apps, video and IPs
  • It analyses sites from a normal Spanish IP, not a datacentre one
  • What one account catches protects the rest
Layer 2 · after the lead

⭐ Lead Rating

Scores every lead from 0 to 100 using what they did before contacting you, and uploads that value to Google Ads as a conversion — with its stages up to the sale.

  • Raw, qualified, opportunity and sale
  • Offline conversions automated, no sheets or schedules
  • A worthless lead is removed from the account
Layer 3 · after the data

🎛️ Automatic Smart Bidding

Moves each campaign's target ROAS and budget every night, within your limits, going after profit instead of volume. Included in the plan and optional: if you'd rather run bids by hand, that's fine.

  • Target ROAS reviewed every day, per campaign
  • Budget per campaign with its cap, and the month's spend under control
  • It undoes on its own whatever makes things worse

Why fraud and quality go together

Fixing only one of the first two is no use

Block the fraud but keep counting every form the same, and the algorithm will keep buying the cheap ones. Score the leads while fake traffic still comes in, and you are scoring bots. Those two don't come apart. The third layer is what puts the result to work by moving the bids, and that one is optional: leave it to us or keep running it yourself.

🩸

What gets taken without showing

Today's fraud isn't a repeated click: it is websites built to receive ads, apps that open themselves, video nobody watches and forms filled in by a program. In Google's report, all of that shows up as one more impression, one more click and one more conversion.

And the money isn't the worst part: Smart Bidding learns from it and goes looking for more of the same.

🎯

What Google cannot know

To Google Ads, a form from someone who bought €8,000 and one from a made-up phone number are exactly the same conversion. Same bid, same value, same weight in the learning.

Nobody inside Google Ads can fix that: whether the lead was any good lives in your business, not in the platform.

What almost nobody has

The pieces that make this actually work

None of them is a marketing headline: they are technical decisions taken because the previous approach didn't work, and each one can be explained.

🌍

An IP that is allowed to look

Farm websites block IPs from Google's datacentres. Analysing them from inside Google Ads is impossible: they simply don't answer. Our analyser comes in from an ordinary Spanish IP, and then you see what's inside.

⚙️

The engine doesn't run in your account

The heavy lifting — measuring, analysing fraud, scoring and uploading conversions — lives on our server, not in a script inside your Google Ads: no 30-minute cap, no quotas, nothing left half-done. Layer 3 is a script in your account, because for moving bids that's an advantage. Why →

🔗

Conversions configure themselves

The four conversion actions — raw, qualified, opportunity and sale — are created in your account through the API, with a dry run first. No configuration screens, no import schedules to set up.

🌙

And they upload every night

No files in Drive, no import schedules and no service accounts to share. That plumbing is exactly where most people's measurement breaks: here it doesn't exist.

⏱️

90 days to upload, 7 to learn

You have 90 days to upload an offline conversion, but the algorithm only learns from what arrives in the first 7. That's why the lead uploads the same day, already carrying its score. Explained here →

🧪

Enhanced conversions

Hashed email and phone so Google can match the conversion even if the click identifier is lost. They only travel if that person accepted advertising, and never in the clear.

🧹

Worthless leads are deleted from Google

Marking a lead as bad isn't an internal note: the conversion is removed from your account, so the algorithm stops learning from it. And if you were wrong, unmark it and it comes back.

👥

Duplicates, handled

The same person writing three times is not three leads. They are grouped with a window you choose, and whatever you mark on a duplicate goes to the original lead, which is the one carrying the click.

🧷

The click survives the subdomain

If your ad lands on landing.yoursite.com and the form lives on yoursite.com, attribution is lost on the way. Our tracking recovers it at the root domain.

From the visit to the bid

All of this happens on its own, every day

You put one tag on your website and mark in a sheet which leads were any good. The rest of the chain runs untouched.

1

Someone clicks your ad

The click identifier is stored, along with where they came from, what they do on the page — and whether they accepted advertising.

🛡️

The Shield decides whether that traffic deserves to keep costing

It scores the visit, cross-checks the site, the IP and the placement against the farm registry, and excludes what isn't an interested person.

2

A form arrives

The ones that aren't leads — account logins, cart, search, newsletter — are discarded, and the rest is grouped with that person's duplicates.

⭐

Lead Rating scores it, 0 to 100

Using what they did before writing, where they came from, what they wrote and what matters in your business. It lands in your sheet already sorted by quality.

3

Your team marks what happened

Qualified, opportunity, sale with its amount — or worthless. One checkbox. With a CRM, your CRM does it and nobody marks anything.

🌙

That night, Google Ads receives it

Each stage uploads as its own conversion, with its value and its date. Worthless ones are withdrawn. Same thing towards ChatGPT Ads, if you use it.

📈

And the algorithm changes target

It stops chasing cheap forms and starts looking for people who resemble the ones who buy from you.

🎛️

And that night bids and budgets are adjusted

Each campaign's real return is compared with the one you asked for, and the target ROAS and the money move, in short steps and within your limits. The third layer →

Not a closed formula

The rating fits your business, and recalibrates every month

A good lead doesn't mean the same thing in a clinic, a driving school and an industrial machinery firm. That's why the score doesn't arrive off the shelf:

  • Your project's signals are added to it: the conversions that really count, the area you operate in, what tells your good customer apart.
  • It recalibrates itself every month with your account's real results: the ranking doesn't stay frozen on installation day.
  • And it audits itself: the report shows you, band by band, how many of the leads it scored high ended up being worth it.

The question almost nobody dares to answer

A scoring system that never publishes its own accuracy is an opinion with decimals. The monthly report includes the calibration curve: for each score band, what share of those leads turned out good.

BandLeadsHit rate
90-100—high
70-89—high
40-69—medium
0-39—low

With your data, and with the cumulative cut-off: "working from 70 up, you capture X % of your good ones".

Personal data

Built so you can defend it in front of a European client

This system handles people's data, so consent isn't a checkbox bolted on at the end: it is part of the design.

✋

Only with permission

Hashed email and phone travel to Google only if that person accepted advertising.

🎚️

And you decide the rest

For those who didn't, you can send nothing — or send only the click identifier: the conversion counts and no personal data travels.

🗓️

Retention and anonymisation

How long data is kept and when it is anonymised is set by you, per account, not by us.

🏥

Sensitive data

If your form may carry health data, the lead can be measured without storing its content.

Your leads are never mixed with anyone else's. Each account has its own store and its own sheet. The only thing shared between clients is the registry of fraudulent sites and farms — which holds no personal data, and is what lets a new client start protected from day one.

Honesty

What Ninja Leads is not

It is not a CRM

It doesn't replace where you manage your leads: it connects to the one you already use, and if you don't use any, the sheet works as an inbox sorted by quality.

It is not a click filter

Blocking IPs is the easy part and the least useful one. What moves the needle is excluding the inventory that produces that traffic and correcting what the algorithm learns.

It doesn't install itself

There is a setup with us: the tag on your site, your measurement reviewed and the rating calibrated against your business. That's why it isn't sold by the unit.

Next step

We look at your account before proposing anything

There is no catalogue price: what it costs depends on how many contactable leads you get each month and how your measurement is set up. Tell us both and we'll tell you what can be done — and what can't.