Lead Scoring for Agencies: Use the Website as the Score

Lead scoring is giving each lead a number that says how likely it is to become a customer, so you contact the best ones first. Most models score what a lead does on your website. An agency that builds its own list has no such data. It can score the prospect's website instead.

What is lead scoring?

Lead scoring is a way to rank leads by a number, so that sales time goes to the leads most likely to buy. The number is built from signals, and each signal is worth some points.

Standard models use two kinds of signal.

A lead scoring model is the list of signals, the points for each one and the cut-off that decides who gets a call. Some teams set the points by hand. Others let software learn them from past deals, which is called predictive scoring.

Why do standard models fail for an agency?

Standard models score behavior, and a cold list has none. The dentist you found on Google Maps has never opened your emails or visited your pricing page. Every behavior signal reads zero.

That leaves fit. Fit narrows a list to dentists in Mesa with 5 to 20 staff. It cannot say which of those dentists needs a new website, so every practice on the list gets the same score.

Predictive scoring does not close the gap. It learns from a long record of past deals, and a small agency has a short one. Our post on AI lead generation covers what the software should and should not decide.

What can an agency score instead?

The prospect's own website. It is public, it is the thing you sell, and it shows need without a single conversation.

A website score turns "does this business need us" into a number. Five signals are enough to start.

The 2.5 seconds is Google's own mark for a good page load, measured at the 75th percentile of visits (web.dev).

One thing runs the other way from a standard model. In a website score, the lowest number is the best lead.

Leadodo scores websites this way, so read our view with that in mind. Twip, the AI inside Leadodo, scores every website in a search out of 100 across 10 domains. They are design, speed, mobile, security, SEO, accessibility, content, conversion, trust and technology. Sites that score below 40 form the shortlist.

Does a website score predict replies?

In our test it did. We scored 20,000 Dutch company websites on 206 indicators, then sent 3,570 cold emails to three groups over 5 working days.

Selecting by score made a business about 3.2 times more likely to reply positively. The full method is in our write-up of the test.

Read the table with three cautions.

1. B against C is not proven. The gap between the generic email and the specific one was not statistically significant, with p = 0.238. The study shows that the score picks better prospects. It does not show that naming the problem adds more.
2. The companies were Dutch and the channel was email. The study says nothing about letters or LinkedIn. A US agency should take the direction from it and measure the size itself.
3. Businesses with no website were left out. A website score cannot rank them.

Scoring was also cheap. The whole experiment cost €242.25. For an earlier campaign, reviewing 1,382 websites by hand would have taken about 184 hours, at about 8 minutes a site.

How do you build a lead scoring model for an agency?

Pick five to ten things you can check on any website, give each a point value and set a cut-off. Six steps cover it.

1. Choose the signals. Start with the five in this post. Add the ones that match what you sell. An SEO agency adds page titles and whether the business shows up for its own trade and town.
2. Give each signal points. Make the points add up to 100, and give the most to the problems you fix best.
3. Score 50 websites in one trade and one town. At the study's estimate of about 8 minutes a site, that is 400 minutes, or a little under 7 hours.
4. Sort from lowest to highest. The lowest scores get your first letters and calls.
5. Set the cut-off once you see the spread. The study took the weakest 20%. A fixed line, such as 40 out of 100, is easier to explain to a colleague.
6. Check the score against replies. After each campaign, compare replies below and above the cut-off. If the two groups reply at the same rate, change the signals.

Here is a hand model applied to one business. Sunridge Dental at sunridgedental.example is made up, and so are the points.

Sunridge Dental scores 30, which is under a cut-off of 40, so it goes on the shortlist. The four zeros also tell you what to write about. This hand model is a teaching example with made-up weights.

What does a lead score not tell you?

Whether this one owner will buy. A score ranks a list. It says nothing about budget, about who decides or about timing.

Those questions belong to lead qualification, which happens one lead at a time, mostly on a call. Score the list first, then qualify the top of it.

A website score also has two blind spots. It cannot rank a business with no site, and our guide on how to find businesses without websites covers those. It also says nothing about where your list came from, which is the subject of B2B lead generation for agencies.

Last, the score is for you. The owner should never see a number. Send one finding, in plain words, and our list of nine signs a business needs a new website shows how to phrase each one.

Questions

What is lead scoring in simple terms?

Lead scoring is ranking your leads by a number so you know whom to contact first. Each lead earns points for signs that it fits what you sell and is ready to buy. The leads past a set cut-off get your time first.

What is a good lead score?

It depends on the direction of the model. In most CRM models a higher score means a better lead. In a website score a lower number means a weaker site and a better lead. Know which way your model runs before you set a cut-off.

What is the difference between lead scoring and lead qualification?

Lead scoring ranks a whole list by a number before you contact anyone. Lead qualification decides whether one lead is worth a sales call, using questions about need, budget, who decides and timing. Score first, then qualify the leads at the top.

Can you do lead scoring without a CRM?

Yes. A spreadsheet with one row per business and one column per signal is enough. Add a total column, sort by it and work from the weakest site upward. A CRM helps once the list is too long to keep current by hand.

Is predictive lead scoring worth it for a small agency?

Rarely. Predictive scoring learns from past deals, and a small agency has too few of them. A rule-based website score needs no history. It only needs the websites, which are public.

---

Next step. Run one search and let Twip score every website in it, with the weakest sites on the shortlist. See how Twip scores a website.