What Is Lead Validation, and Why Half Your Conversions Are Not Leads
A conversion is not a lead. Until someone separates the two, every cost per lead, channel comparison and scoring model you build is running on inflated numbers.
Most marketing teams optimize against a single number: conversions. Forms submitted, calls tracked, goals completed.
It is the number in the dashboard, so it becomes the number that decides budget. But it is only as good as the event tracking behind it, and event tracking is built to count actions, not to judge them.
A meaningful share of those conversions could never have become revenue. Job applicants, support requests, spam submissions, wrong numbers and vendors pitching you all land in the same bucket as genuine buyers.
Lead validation is the fix. It means reviewing every form submission and tracked call and marking it as a sales lead or not, so the number you optimize against is the number that actually matters.
The short version: validate first, then score. Validation removes the inquiries that were never commercial, and lead scoring ranks the ones that are.
Doing the second without the first means your scoring model, your routing rules and your channel budgets are all learning from noise.
How much of your conversion data is actually noise
The most widely cited benchmark comes from Straight North’s “The Critical Importance of Lead Validation” study, which reviewed more than 373,000 inquiries generated by client websites. Roughly half turned out to be something other than a sales lead.
Two caveats, because this figure gets repeated without them. The data is agency-side and skews toward SMB lead generation running SEO, PPC and display, so your own ratio will differ.
It is also a decade old, and both bot traffic and form spam have changed since. The number to trust is the one you produce from your own sample, not the one from the study.
That does not make the study useless. It establishes that the gap is large enough to change decisions, which is the only part you need.
Forms and calls also fail differently. Submissions tend to fail on completeness and spam, while tracked calls fail on intent, with misdials, telemarketers and support calls making up much of what call tracking counts as a conversion.
What counts as a non-sales conversion
A non-sales conversion is any inquiry that registers in analytics but has no realistic path to revenue. Naming the categories matters, because each one has a different fix and points at a different part of your funnel.
| Category | What it looks like | Where the fix usually lives |
|---|---|---|
| Bot and spam submissions | Gibberish fields, disposable email domains, high submission velocity from one source | Form hardening, bot protection, honeypot fields |
| Incomplete submissions | No usable phone or email, placeholder text, abandoned multi-step forms | Field validation, progressive profiling to cut form friction |
| Vendor solicitations | Agencies and suppliers pitching you through your own contact form | Lead routing rules, keyword filters on message body |
| Job applicants | Career enquiries submitted through the sales or demo form | Clearer careers navigation, form segmentation |
| Existing customer support | Account issues arriving through marketing forms and tracked numbers | Dedupe against existing Salesforce accounts, support routing |
| Misdials and robocalls | Sub-15-second calls, auto-dialers, menu loops, voicemail dead ends | Call disposition codes, minimum duration thresholds |
Three to five categories is enough to start. The point is not taxonomic completeness, it is being able to say “this channel produces twice the spam of that one” and act on it.
The timing finding, and what it does not mean
The same study reported that around 85 percent of sales leads convert on their first visit to the site, 10 percent on the second and 3 percent on the third.
This gets quoted as evidence that nurture does not matter. That reading is wrong, and the reason is worth understanding.
The measurement is visit-based conversion tracking on a single domain. A buyer who researches you for six weeks, then clears cookies or switches from mobile to desktop before filling in a form, is recorded as a first-visit conversion.
That is an artefact of how first-touch attribution draws a session boundary, not a finding about buyer patience. It tells you something real about the moment of submission, and very little about the research that preceded it.
So the implication is not “stop nurturing.” It is that the anonymous research before a form fill is largely invisible to standard conversion tracking.
Persistent website visitor tracking and a multi-touch attribution view show a very different picture of the same journey. They read the digital body language that precedes a submission, not just the touchpoint that closes it.
Until you have that, assume any page can be the only page your analytics ever credits. Make it capable of converting on its own.
A validation workflow you can actually run
Automated checks at the point of capture, plus a small amount of human review for edge cases. Calls almost always need the human layer.
Sample before you build anything
Pull the last 50 to 100 inquiries and classify each one by hand. It takes about an hour and tells you which categories dominate.
Building filters before this usually means solving the wrong problem carefully.
Automate the obvious rejections
Required fields, email format checks, blocked disposable domains, bot protection, duplicate detection against existing CRM records.
These catch the biggest buckets from step one with no ongoing effort.
Route ambiguous cases to review, not to sales
Set up a queue for anything a rule cannot confidently classify. Sending an uncertain inquiry straight to a rep costs more than the two minutes of review.
A rep who works three junk leads in a row stops trusting the source entirely, and that erodes sales qualification faster than any scoring error.
Classify calls by disposition, then feed it back
Connected sales conversation, support request, misdial, spam. Push the outcome into campaign attribution reporting so call-heavy channels are judged on the same basis as form-heavy ones.
Version your rules, so you can compare performance before and after a change without rewriting history.
What this changes in reporting. Calculate cost per lead using validated leads as the denominator, not raw conversions, at the level you actually optimize: campaign, ad group, keyword.
Run “cost per conversion” and “cost per validated lead” side by side for a month. The gap is usually the most persuasive slide you will build all quarter, and it frequently reverses which channel looks like the winner.
If you report inside Salesforce, B2B marketing attribution and campaign reporting is where the validated count needs to land, so the same denominator carries through to pipeline.
Validation cleans the list. Behavior tells you who to call first.
SalesWings is not a lead validation tool. It does not review call recordings or filter spam submissions.
It solves the problem that sits immediately after validation: once you know which inquiries are real, which of them deserves a rep’s next hour.
That matters more once validation is in place, not less. A clean list is still an undifferentiated list.
Behavioral data collection captures what leads and accounts actually do on your site, portal and campaigns. That becomes Salesforce lead scoring living natively on your objects, alongside fit and firmographic data.
- Buying signals that reveal intent before a form is submitted, including the research the timing data above cannot see
- Deterministic, rule-based behavioural lead scoring your team can explain to sales, not a score nobody can interrogate
- Fit scoring and lead grading on any Salesforce data you hold, so an ICP mismatch does not get worked just because it was active
- Real-time sales alerts when a validated lead returns and revisits high-intent pages
There is a useful feedback loop here too. Behavioral data catches a category validation often misses: the technically complete, non-spam submission from someone with no commercial interest.
Someone who fills in a demo form and never touches a product or pricing page has an engagement score that looks nothing like someone who read four pages first.
See what your leads did before they filled in the form
Behavioral lead scoring, fit scoring and real-time alerts, native to Salesforce.
For the practical implementation, see Salesforce Lead Scoring.
Frequently asked questions
On this page
- Why should I care about lead validation as a marketer?
- How does lead validation improve lead prioritization?
- How do I make cost per lead accurate?
- How do sales and marketing stay aligned on this?
- Why does conversion attribution break for leads?
- Does validation have to be manual?
- Is lead validation the same as lead scoring?
Why should I care about lead validation as a marketer?
Because a large share of what your dashboard calls a conversion is not a sales opportunity, and that distorts every optimization decision downstream.
If your reporting treats spam, empty forms, job seekers and wrong numbers as leads, you will over-credit the channels that produce the most noise. You keep funding what is not creating pipeline.
Validation lets you see true performance by campaign, keyword and landing page. It also protects your sales team’s time, which tightens the feedback loop on creative, offers and targeting.
Start by sampling 50 to 100 recent inquiries and classifying them by hand, then automate the rules that catch the biggest buckets.
How does lead validation improve lead prioritization?
It ensures your scoring and routing models learn from real sales leads rather than from noise.
If a significant share of your leads are non-sales, your lead score thresholds, alerts and SLAs get tuned to the wrong signals. Reps chase inquiries that were never going to close while genuine buyers wait.
Once validated, you can rank leads on a cleaner intent score and fit score, then measure lift by segment. This matters most in high-volume funnels, where small gains in quality become large gains in capacity.
Pair validation with multi-layer scoring and clear routing bands. The high-volume lead prioritization playbook works as a checklist.
How do I make cost per lead accurate?
Divide spend by validated sales leads, not by raw form fills or total tracked calls.
Unify every inquiry source into one table with campaign metadata attached: source, medium, ad group, keyword, landing page. Then validate each inquiry, so the denominator reflects real demand.
Calculate cost per lead at the level you actually optimize and track it weekly, which also catches tracking breaks fast. This regularly changes which channels look like winners, because it corrects for spam-heavy sources.
How do sales and marketing stay aligned on this?
One shared definition of a good lead, one agreed handoff moment, one feedback loop in the CRM.
Agree the qualification criteria covering both fit and intent, and the exact point at which an MQL becomes an SQL. Sales also needs context at handoff: campaign, page path, key engagement signals.
Implement it with routing rules, SLAs and a simple disposition taxonomy, so sales can reject or recycle leads with reasons marketing can act on. A shared sales and marketing alignment layer in the CRM is what makes that agreement survive a quarter.
Without that last part, validation stays a marketing-only exercise. Sales keeps its own private opinion of lead quality.
Run a quarterly reset as your ICP and channels evolve. The sales and marketing alignment guide covers the operating model in more detail.
Why does conversion attribution break for leads?
Because standard tracking records a thank-you page event and nothing about quality, identity or outcome.
Goal tracking cannot tell you whether a submission had usable contact details, matched your ICP, or ever reached a human. So it treats every conversion as equal.
It gets messier when calls are not stitched to campaigns, when consent flows and redirects drop parameters, and when people switch devices mid-research. That last one is where website visitor identification does work analytics cannot.
Lead tracking improves on this by tying each inquiry to its marketing source and carrying that record forward into CRM stages. Revenue attribution then closes the loop rather than stopping at the form.
Fix the largest leakage point first. Usually that is call capture, parameter persistence, or handoff consistency into the CRM.
Does validation have to be manual?
No, but the edge cases usually do, and calls almost always do.
Automated rules handle most clear-cut cases: format checks, blocked domains, duplicates, bot detection. For most teams this sits with marketing operations to build and with sales to maintain.
What rules struggle with is intent expressed in free text, which is why a review queue for ambiguous submissions earns its keep. Phone calls are the harder half, since the only reliable signal is what was said.
Is lead validation the same as lead scoring?
No. Validation is a yes or no filter. Scoring is a ranking applied to whatever survives it.
Validation answers “could this ever become revenue,” and the answer is binary. Scoring answers “of the ones that could, who is closest,” and the answer is a range that feeds lead qualification and a judgment about sales readiness.
Scoring without validation produces confident-looking rankings of a population that includes inquiries with no commercial potential at all. Validation without scoring gives you a clean list with no order to work it in.
Source: The Critical Importance of Lead Validation in Internet Marketing, Straight North. This article was originally published as a guest contribution by Aaron Wittersheim, Chief Operating Officer at Straight North, and has been rewritten and updated by the SalesWings team.
