In a B2B company, the sales funnel usually looks fine on a slide: inquiries at the top, orders at the bottom, a few CRM stages in between. The problem is that money is not lost at the stages, but between them. An inquiry waits a day for a reply, the quote takes a week to prepare, and once it is sent nobody calls. None of these moments looks like a loss, so nobody counts them.
In this article we show the five places where a B2B sales funnel most often leaks, what it costs and what can be automated at each one without replacing your whole system. We write as a company that works with manufacturers on quoting systems, configurators and sales automation, so we see these leaks up close. Halfway through the article there is a calculator: enter your own numbers and see at which stage you lose the most.
What a B2B sales funnel is and how it differs from an online shop funnel
A sales funnel is the path a customer takes from first contact to order, described as a series of stages with the number of people at each one. In an online shop this path takes minutes and runs through the website, the basket and payment. In B2B sales, especially at manufacturers, it takes weeks or months, and there is almost always a person in the middle who prepares the quote.
A typical B2B sales funnel has five stages:
- Inquiry from a form, email, phone call, trade fair or distributor.
- Qualification, meaning checking whether this is a real customer, what they need and by when.
- Offer with pricing, often calculated by hand by a salesperson or a process engineer.
- Follow-up and negotiation: questions, revisions, discounts, the decision on the customer's side.
- Order and handover to production or delivery.
The difference from an online shop matters for automation. In B2B the point is not to replace the salesperson, but to stop wasting time on things that do not need their expertise, and to make sure no inquiry gets stuck along the way. Each of the five stages has its typical leak.
How to measure your funnel before you start fixing it
Most companies know two numbers: how many inquiries came in and how many orders there were. That is not enough to know what to fix. Five extra measurements are enough, and you can collect them even from your inbox and Excel:
| Stage | What to measure | Warning sign |
|---|---|---|
| Inquiry | Time from arrival to the first human reply | More than one working day |
| Qualification | How many inquiries drop out and why | Nobody knows why they drop out |
| Offer | Time from inquiry to sending the quote | Several days for products that repeat |
| Follow-up | What share of quotes has a next contact scheduled | Quotes hang without a date for the next step |
| Order | Quote win rate and average order value | Win rate falls even though inquiry numbers stay the same |
The same data is useful later for planning. If you want to use it to work out how much you will sell in the coming months, we covered that in our article on forecasting sales from the funnel. Here we focus on something else: where those numbers slip away.
5 places where the sales funnel loses inquiries
1. The first response comes too late
A customer who sends an inquiry to a manufacturer rarely sends it to just one company. They usually write to three or four and talk first to the one that gets back to them first. If your reply arrives the next day, the first conversation with a competitor has often already happened, and you start by playing catch-up.
The most common cause is not lazy salespeople, but the route the inquiry takes. The form lands in a shared inbox, someone forwards it, the salesperson is on the road, and a weekend inquiry waits until Tuesday. Nobody notices the inquiry is sitting there until the customer calls to complain.
What to automate:
- A reply with real information within a minute, not just "thank you for your message". For example: who will handle the inquiry, by when they will come back with a quote and what other details we still need.
- Routing the inquiry to a salesperson by region, product or order size, instead of forwarding it by hand.
- Reminder and escalation: if nobody has replied after 4 hours, the inquiry goes to the sales manager.
- A form that qualifies straight away: a few questions about quantity, deadline and application save one round of emails.
This is the cheapest fix in the whole funnel. A well set up form, rules in the CRM or a simple email integration is usually enough, without a new system.
2. The quote takes days
At manufacturers this is usually the biggest leak. The quote is built in an Excel sheet with dozens of columns, material prices have to be checked in a separate file, and for a non-standard size the salesperson has to ask a process engineer. A single quote takes anything from an hour to several days, and in the meantime the customer gets a quote from someone else.
The second cost is less visible: your best salespeople spend their time calculating instead of talking to customers. The more inquiries you get, the longer each quote waits, so growing demand paradoxically lowers your win rate.
What to automate: pricing. The rules that today live in the salesperson's head and in Excel can be captured in a configurator or quoting system that calculates the price, checks whether a given variant can be manufactured and generates a PDF. In one of our projects for a manufacturer, preparing a quote went from 2 days to 30 minutes, and at an accounting firm from 45 to 6 minutes. If you want to check whether it pays off, see how to calculate the return on a configurator.
A configurator available to customers on your website also changes the top of the funnel. Customers check variants and an indicative price themselves, and the salesperson receives an inquiry with a ready specification rather than "please send me a quote".
3. Salespeople answer the same questions over and over
Is it in stock? What is the maximum width? When will you ship the order? At one company we worked with, about 150 questions like these came in every week, and 60% of them were repeated almost word for word. Each one takes a few minutes on its own, but together they eat up several days of work a month.
This is a different kind of leak: you do not lose the customer, you lose time the salesperson could spend on quotes and conversations with larger customers.
What to automate: answers that already exist somewhere. First, a knowledge base with answers to the most common questions, available to customers and to the team. Then an AI assistant that answers based on catalogues, manuals and order data, and passes harder questions to a person. Which salesperson tasks can be taken over this way, we described in our article on AI agents in B2B sales.
4. Quote sent, then silence
This is the most frustrating leak, because all the work has already been done. The quote went out, the customer did not reject it, but did not reply either. The salesperson has dozens of open quotes in their head and calls the ones they remember. The rest quietly drop out of the funnel after a few weeks, often because the customer had one question and nobody to ask.
What to automate:
- A reminder sequence: after 2 days a short message asking whether the quote is clear, after 5 days a "call" task in the CRM, after 10 days a message with extra information, such as the delivery date.
- An online quote instead of a PDF attachment. The salesperson gets a notification when the customer opens the quote and calls while the customer is thinking about it.
- A next-step date on every quote. The system does not let anyone leave a quote without a scheduled contact.
Note: automatic reminders must sound like a message from a person and must stop once the customer replies. A series of three identical emails does more harm than no follow-up at all. If the CRM in your company does not work the way it should, read why CRM implementations fail.
5. All inquiries are treated the same
An inquiry for one unit and an inquiry for 200 units land in the same queue and are handled in order of arrival. A salesperson can judge them by eye, but only when they have time to read them. As a result, a large, urgent inquiry waits behind several small ones.
What to automate: lead scoring. You do not need artificial intelligence for this; the data the customer provides anyway is enough: the value calculated in the configurator, quantity, deadline, industry, whether they are a regular customer. The system adds points and flags the inquiries that need to be handled first. It pays off most when the configurator is connected to the CRM, because the specification and value reach the salesperson without retyping.
Calculator: how much you lose at each stage of the funnel
Below you can check what these five leaks look like in your company. Enter figures from recent months, or use estimates if you do not have them. The calculator will show how much margin and time you lose per year at each stage, rank the stages from the biggest loss and suggest what to automate first.
Sales funnel leak calculator
Enter a few numbers about your sales. The result updates as you type and shows how much margin you lose per year at each of the five funnel stages and where to start with automation.
How to read the result? We calculate sales losses on margin, not revenue, because only margin actually stays in the company. Time spent on repeated questions is calculated at labour cost. The "recoverable" amount is 60% of losses, because no automation closes a leak completely. All the assumptions are listed below the result, so you can change them in your own calculation or show them to your board together with the numbers.
The calculator is deliberately cautious. It does not include customers who never send an inquiry at all because they cannot check the price themselves, nor the fact that a faster quote often lets you hold a higher price. In practice, the gain from fixing the funnel is often bigger than the result.
What to automate first: an order that pays off
The calculator shows where you lose the most, but the order of implementation also depends on cost and time. At manufacturers, this order usually works best:
| Step | Automation | Project size | When you see the effect |
|---|---|---|---|
| 1 | Measurement: response time, quote time, quote statuses in one place | Small, the current CRM is often enough | After 2-4 weeks, once the first data comes in |
| 2 | A reply with real information, inquiry routing, escalation | Small | Immediately |
| 3 | Follow-up sequence and an online quote with an open notification | Small to medium | After the first month |
| 4 | Configurator or quoting system | Medium to large, an MVP usually takes about 8 weeks | After go-live, the biggest effect in the whole funnel |
| 5 | Lead scoring and an AI assistant on company documents | Average | Once you have data from steps 1-4 |
Steps 1-3 can be done in a month and usually do not need a new system. They deliver a quick effect and, more importantly, data that makes it easier to justify a bigger project, such as a configurator. For a broader overview of what can be automated in the sales department beyond the funnel itself, see our articles 7 processes to automate in the sales department and B2B sales automation in 2026.
What not to automate in the funnel
Automation should take load off the salesperson, not keep them away from the customer. There are places where it is better to keep a person involved:
- Conversations about large and non-standard orders. Here the customer is also buying trust in a person who understands their problem.
- Negotiation and discounts. The system can suggest ranges, but the decision should stay with the salesperson.
- Responses to complaints and problems. An automatic reply to a complaint usually makes things worse.
- A chatbot without access to company data. A bot that answers in generalities or makes up specifications costs more trust than it saves time.
A good rule: automate what is repetitive and has one correct answer. Where the answer is "it depends", keep a person and give them better data.
Frequently Asked Questions
What is a B2B sales funnel?
It describes the path a business customer takes from the first inquiry to the order, split into stages: inquiry, qualification, quote, follow-up with negotiation, and order. At each stage you measure how many inquiries move on and how many drop out. In B2B the funnel is longer than in an online shop, because several people make the decision and the quote is usually prepared by a person.
How do you check where your sales funnel is losing customers?
Measure five things: time to first reply, the share of inquiries that drop out before a quote, quote preparation time, the share of quotes with a scheduled follow-up, and quote win rate. The stage with the worst result relative to order value is your biggest leak. You can estimate it for your company with the calculator in this article.
What should you automate in the sales funnel first?
Usually measurement of response times and quote statuses, then an automatic reply with routing to a salesperson and a follow-up sequence. These steps are cheap and deliver results within a month. The biggest gain for manufacturers comes from automating pricing in a configurator or quoting system, but that is a bigger project.
Do you need a CRM to automate the funnel?
Not always for the first steps. A reply with real information, inquiry routing and reminders can be set up with a form and email. A CRM becomes necessary when you want to see all quotes in one place, plan follow-ups and score leads. More important than choosing a system is describing the process and appointing someone who is responsible for it.
Doesn't automatic follow-up put customers off?
It does if it is a series of identical emails that keep coming even after the customer has replied. Good follow-up is short, adds new information (a delivery date, an answer to a common question, an example project) and stops when the customer writes back. What works best is combining an automatic message with a call from the salesperson at the right moment.
Next step
If the calculator revealed a leak that surprised you, send us the result or a short description of how an inquiry moves through your company today. We will tell you which automations will deliver results in the first month and which are worth planning as a bigger project. You will find examples of working configurators and quoting systems in our showroom.
Write to us or see why customers increasingly look for a configurator before they send an inquiry.
Read also
- B2B lead generation in 2026: channels, costs, what no longer works
- Sales Forecasting: How to Stop Guessing and Calculate Your Funnel Forecast
- Six Sales Signs That Are Costing Your Company Real Money
- ROI from the configurator: how to realistically calculate the return on investment
- The best CRM for a manufacturing company in 2026: a ranking of 29 tools
