A lead form submission is not automatically a sales opportunity. For a London business spending thousands on Google Ads, Meta Ads or Microsoft Ads, treating every enquiry as equal can make campaign reporting look healthier than the pipeline really is. The decision to set lead scoring thresholds gives marketing and sales a shared definition of what deserves attention, what needs more nurturing and what is unlikely to justify further budget.
The aim is not to create an elaborate scoring model for its own sake. It is to connect paid media activity to the outcomes that matter commercially: qualified conversations, credible opportunities and revenue. A useful threshold helps reveal which campaigns produce better leads, not merely lower cost-per-lead figures.
What lead scoring thresholds should achieve
Lead scoring assigns a value to signals that indicate fit and intent. A threshold is the score at which a lead changes status, such as becoming a marketing-qualified lead (MQL), a sales-qualified lead (SQL), or entering a nurture sequence.
For example, a lead who matches your target sector, works for a business of the right size and requests a consultation is likely to be more valuable than someone who downloads a general guide using a personal email address. Scoring turns that difference into a consistent operational decision.
A well-set threshold should do three things. It should prioritise follow-up for the leads most likely to progress, prevent sales teams wasting time on poor-fit enquiries, and give paid media reporting a more reliable quality measure. If a campaign generates 80 leads but only two cross the qualification threshold, it needs a different assessment from one that produces 25 leads and ten credible sales conversations.
The right number is not universal. A high-value B2B service with a long buying cycle may accept fewer leads at a higher score. A business with a short sales process and high enquiry volume may need a lower threshold, provided its sales team can respond quickly and qualify efficiently.
Start with evidence from your existing pipeline
The most common mistake is choosing a score because it feels sensible. A threshold of 50 points means very little unless it corresponds with actual sales outcomes.
Begin by looking at closed-won customers, genuine sales opportunities and leads that went nowhere. Review the information available at the point of enquiry: source, campaign, search term or audience, job title, company size, location, service requested, form answers, pages viewed and actions taken after the first visit.
Look for patterns rather than isolated anecdotes. Do closed-won customers tend to request a specific service? Are they usually within a particular turnover range? Do they engage with pricing or case-study content before speaking to sales? Do leads from a certain campaign produce meetings but rarely progress beyond them?
This exercise also exposes tracking gaps. If your CRM records whether a lead became an opportunity but does not retain the original campaign source, it becomes difficult to make confident budget decisions. Likewise, if every form fill is sent to Google Ads or Meta Ads as the same conversion, platform optimisation will learn from quantity rather than quality.
Before refining thresholds, make sure the journey from advert click to CRM outcome is clear enough to trust. Perfect attribution is rarely realistic, but a consistent view of lead source, qualification status and revenue is a practical starting point.
Separate fit from intent
The most dependable scoring models use two types of signals: fit and intent. Fit measures whether the lead resembles a customer you can serve profitably. Intent measures whether that person appears ready to have a meaningful commercial conversation.
Fit criteria might include UK location, relevant industry, company size, decision-making seniority or stated budget. Intent criteria can include booking a consultation, submitting a high-intent form, returning to the site, viewing service pages, attending a webinar or responding to follow-up.
Neither category is sufficient alone. A senior decision-maker at an ideal company may simply be researching. A highly engaged visitor may be a student, competitor or business outside your service area. Combining the two reduces false positives.
Use negative scores as well. A personal email address is not automatically a bad lead, particularly for some services, but it may deserve fewer points than a verified company domain. Enquiries outside your geographic coverage, irrelevant job roles, implausible phone numbers and clearly unsuitable requirements should reduce a score or trigger exclusion.
How to set lead scoring thresholds from real outcomes
Create a simple points model first. Avoid assigning tiny values to dozens of behaviours. It is better to score the signals your sales team genuinely recognises than to build a system nobody can explain.
A professional-services business might award points for a target industry, a company of sufficient size, a relevant decision-maker and a consultation request. It might deduct points for an overseas enquiry where it only serves UK clients, a student address or a request that sits outside its offer. The exact weighting depends on what previous customers have in common.
Then score a sample of historic leads. Include leads that became customers, leads that became valid opportunities but did not close, and leads that sales disqualified. Arrange the scores from highest to lowest and assess where the groups begin to separate.
If most closed-won leads score between 65 and 90, while most disqualified leads are below 35, a provisional MQL threshold around 45 or 50 may be reasonable. A higher score, perhaps 65, could define an SQL requiring immediate sales contact. These figures are examples, not a prescription. The useful threshold is where lead quality improves without starving the sales team of viable opportunities.
Do not set the MQL level so high that only obvious buyers qualify. Marketing can then appear very efficient while potential customers who need modest nurturing are ignored. Equally, a threshold that allows nearly every enquiry through offers little protection against wasted follow-up time.
Use response capacity as a practical constraint
Thresholds need to reflect operational reality. If your sales team can properly contact and qualify 20 leads a week, passing 70 leads through as MQLs is not a sign of successful demand generation. It creates slow response times and lost intent.
This is particularly relevant for paid search, where a prospect may be contacting several providers within an hour. High-scoring leads should be routed quickly, with a clear owner and agreed response standard. Lower-scoring but potentially relevant leads can enter a structured nurture process rather than being left untouched.
Marketing and sales should agree what happens at each score band. For instance, a sales-ready lead may require contact within one working hour, a mid-score lead may receive qualification questions and relevant content, while low-score leads are suppressed or reviewed periodically. The process matters as much as the number.
Feed qualified outcomes back into paid media
Lead scoring becomes commercially useful when it changes how advertising is managed. If your CRM can pass qualified lead and opportunity events back to advertising platforms, campaign optimisation can move beyond raw form submissions.
This does not mean every offline conversion should be uploaded without scrutiny. Use consistent definitions, deduplicate records and allow enough time for leads to be qualified. A rushed or unreliable feedback loop can train campaigns towards the wrong behaviours.
Review performance by qualified-lead rate alongside cost per qualified lead, opportunity rate and, where sales cycles allow, revenue. A campaign with a higher initial cost per lead may be more efficient if a larger share reaches the threshold. Conversely, an inexpensive campaign that repeatedly generates low-score enquiries is often what is wasting budget.
Search terms, audience segments, placements and creative messages all become easier to assess through this lens. A broad keyword may produce volume but poor-fit leads. A specific service-led advert may cost more per click yet produce enquiries with clearer intent. Scoring gives those trade-offs a basis in evidence.
Review thresholds without moving the goalposts
A scoring model should be reviewed, but not constantly rewritten. Check it monthly when lead volume is high, or quarterly for lower-volume businesses. Compare score bands against sales acceptance, opportunities and closed revenue. Ask whether high-scoring leads are actually progressing, whether good prospects are being missed, and whether any criterion is adding noise.
Changes in targeting, offers, seasonality or sales process can alter what a good lead looks like. A new campaign aimed at a different market may need its own rules rather than forcing every prospect into one scorecard. Keep a record of changes so that reporting remains interpretable.
The most useful lead scoring threshold is not the most complicated one. It is the one your teams apply consistently, your CRM can report clearly and your paid media activity can learn from. Start with the evidence already in your pipeline, make the next decision clearer, then refine the model as better lead and revenue data becomes available.
