Lead Management Playbook

Lead management is the agreed path from a first contact to a sales conversation. Get it right and no good lead waits, and sales trusts what marketing sends. It does not take expensive software, only agreed definitions, a simple score and a handoff with deadlines.

Early-stage companies often lose leads in the gap between marketing and sales: a form fill no one follows up for two days, a lead passed to sales that sales thinks is junk, a promising contact who went quiet and was never contacted again.

The research backs a joined-up approach. A study of 337 companies in the European Union identified five ways marketing and sales are organized, and its findings suggest that “the most successful configurations are characterized by strong structural linkages between marketing and sales and a high extent of market knowledge in marketing”^[1]. In a startup those linkages are simple, and the seven steps below build them.

There is no official standard for lead stages. Figure 1 shows seven common ones, each with an entry rule and one owner; keep the number small.

If you use Google Analytics 4, its recommended lead events give you a matching vocabulary for your website and CRM: generate_lead, qualify_lead, disqualify_lead, working_lead, close_convert_lead and close_unconvert_lead ^[2]. The tracking setup post covers how to send them.

Score two things separately. Fit asks whether this person and company match your ideal customer profile. Behavior asks whether they are showing interest now. Put the two together and every lead lands in one of four cells, each with its own next action.

Start simple. A systematic review of 44 studies found that traditional scoring “primarily relies on the experience and knowledge of salespeople and marketers”, and that predictive models, most often decision trees and logistic regression, are expected to replace them because they improve sales performance^[3]. Predictive models need data you may not have yet. One recent study trained its model on 16,600 cleaned lead records, four years of data from a B2B software company’s CRM^[4].

So, in practice:

1. Fit score. Three to five criteria from your ideal customer profile (industry, size, role, region), each worth a few points. The ICP Fit Scorer can do this part.
2. Behavior score. Points for actions that signal buying interest: a pricing-page visit, a demo request, attending a webinar. Give fewer points to actions that signal curiosity, such as reading a blog post, and let behavior points decay over time.
3. Threshold. Set the MQL threshold so sales gets about the number of leads your funnel needs (Step 7) and no more than it can work, then adjust.
4. Recalibrate every quarter. Compare scores with the deals you won and lost. My rule of thumb, not a published threshold: once you have a few hundred closed deals, won and lost, consider a predictive model.

What HubSpot’s annual report shows

HubSpot sells marketing and sales software, and its annual report for fiscal 2014 describes both its own lead engine and its lead-scoring feature in plain terms. This is what the filing says, not an account of how the company made its choices internally. Figure 3 quotes what it says about its own lead engine.

The same filing describes lead scoring in the terms this step uses. Under “Lead Scoring” it lists: “Create a custom lead score based on the behavior and attributes of a potential customer, such as visiting a specific web page, watching certain videos, opening certain emails, having a certain job title, or other custom data in the inbound database. Define which leads are sent to the CRM system for sales engagement based on these criteria”^[5].

In my view, the lesson for a startup is not the volume. It is that one written rule, based on behavior and attributes, decides which leads reach sales.

The best-known evidence is a 2011 article by Oldroyd, McElheran and Elkington^[6]. It reports two analyses of US companies, shown in Figure 4: how fast companies answered a web lead, and how much a fast reply raised the odds of qualifying it, which the authors define as having a meaningful conversation with a key decision maker. Only 13 of the 42 companies in the second analysis sold to businesses; the other 29 sold to consumers.

The article is from 2011 and was published in Harvard Business Review, not a peer-reviewed journal, and I have not found a peer-reviewed study that repeats it. But the direction is clear, and answering fast costs little.

Do it automatically, with rules checked in order. The common ones:

1. Existing account or named target account goes to its account owner.
2. Segment or territory (region, company size, product line) goes to the rep for it.
3. Everything else goes round-robin among available reps.

Set a fallback: if the owner does not act within the SLA time, the lead moves to a manager or the next rep. Log the reason for every reassignment.

A service-level agreement (SLA) says who commits to what, in numbers you can check. A simple one covers:

• Marketing commits to: a volume of MQLs per month by segment, a definition of MQL, and the information passed with each lead (source, pages viewed, form answers).
• Sales commits to: a first-response time for each kind of lead, a number of follow-up attempts over a set period, and a status update in the CRM for every lead.
• Both commit to: reason codes for recycled and disqualified leads, and a weekly or fortnightly review of conversion between stages.

The numbers are yours to set; the point is that both teams sign them. The Lead Scoring and SLA Sheet includes an SLA page and a tracker.

Most B2B purchases involve several people, and they often disagree. In Gartner’s survey of 632 B2B buyers, conducted in August–September 2024, 74% of buying teams showed “unhealthy conflict” during the decision, and teams that reached consensus were 2.5 times more likely to report a high-quality deal^[7]. The same survey found 73% of buyers actively avoid suppliers that send irrelevant outreach^[8].

Three changes follow:

• Link leads to accounts in your CRM, so people from one company show up together.
• Score the account as well as the person. Several engaged people from one target account is a stronger signal than one very active individual.
• Make outreach relevant or skip it. Irrelevant follow-up can cost you the account.

Leads that are not ready need content matched to where they are. A single-company case study of B2B content marketing showed how marketing automation can generate high-quality sales leads through behavioral targeting and personalized content^[9]. A 2024 experiment with 444 US B2B professionals found engagement was higher when content fit the buyer’s stage, but also that “there are no universally correct sequences for presenting content”^[10].

In practice:

• Every recycled lead gets a reason code (no budget, wrong timing, no need yet) and a nurture track to match.
• Behavior that crosses the threshold again sends the lead back to sales automatically.
• Review recycled leads every quarter; some “not now” leads become ready later.

Then measure five numbers, and make them the agenda of the review your SLA sets (Step 5):

1. Conversion rate between each stage, by source and segment.
2. Speed to lead: median time from form fill to first response.
3. SLA compliance: share of leads worked within the agreed time.
4. Pipeline and revenue by source, not just lead volume.
5. Reasons for recycling and disqualification, to improve targeting.

The Pipeline Forecast and Coverage Calculator turns your stage conversion rates into a forecast.

What changes by stage?

What mistakes should I avoid?

• Counting MQLs as success. Measure pipeline and revenue by source (Step 7).
• Scoring blog reads like demo requests. Weight actions by buying intent (Step 2).
• Slow follow-up. Answer within the hour (Step 3).
• No reason codes. Without them you cannot learn why leads fail (Steps 5 and 7).
• Treating people, not accounts, as the unit. Buying groups decide, so score the account too (Step 6).
• A scoring model no one revisits. Recalibrate it against won and lost deals every quarter (Step 2).