A tip is worth printing only when you can say what it changes and by how much. For a small B2B team, four numbers decide the month: how long a new enquiry waits, how many arrive outside working hours, how many go cold after one unanswered question, and what each of those minutes of silence costs you. Measure those, and the rest of the advice on the internet sorts itself out.
This article walks through the four measurements, shows the arithmetic using Asyntai's own published plan figures, and ends with the one change that costs nothing: answering the question the visitor already asked on the page.
The standard list of b2b sales tips has a recognisable shape. Reply faster. Personalise. Follow up. Qualify harder. Close the loop. Each item is true and none of them is useful, because none of them tells you what to change on Monday morning or what the change is worth. A tip that says "reply faster" is a tip that says nothing unless you can put a number on how slow you are, a number on how much speed matters, and a price on the enquiries you lose while you wait.
Here are four things a small B2B sales team can measure this month, with the arithmetic shown in the open and no invented market figures to compare against.
1. How long an enquiry waits before a human replies
The first number is the average wait, and the second is the spread. Average hides the enquiries that waited six hours; the spread finds them. You need both.
Pull your last 200 inbound enquiries. For each one, note the time it arrived and the time a human sent the first reply. Most small teams have this in their inbox or CRM already, even if it takes an afternoon to extract. Once you have the data, sort the enquiries by wait time. The median tells you what a typical buyer experiences. The 90th percentile tells you what your worst one in ten experiences, which is the experience the buyers who went quiet remember.
There is no published figure for what a buyer expects, so do not budget against one. Use your own two numbers instead: the median you measured, and the median you would accept if you were the buyer. The gap between them is the work. The 90th percentile is where deals die, because the slowest replies cluster around the enquiries that needed more thinking, and more thinking often means a bigger contract.
The fastest way to move the median is to stop doing work that does not need a human. A surprising share of first replies on B2B sites are answers to questions already printed on the pricing page, the integration list, the security page or the FAQ. Those replies do not require a person. They require a page that answers the question that was asked.
2. How many enquiries arrive outside working hours
The second number is the share of enquiries that arrived when nobody was looking. This is the number most small B2B teams have never measured, because the inboxes look full during the day and nobody audits what came in at 9pm.
Take the same 200 enquiries. Bin each one by the hour it arrived. Count the ones that arrived outside your published business hours, including the hour before you start and the hour after you stop, because a buyer who sends a question at 4:55pm and gets silence until 10am the next morning has the same experience as a buyer who wrote at midnight.
Whatever the share turns out to be, it is the size of the queue nobody is watching. A buyer who writes at 8pm and gets a reply at 11am the next morning has done a lot of thinking in those fifteen hours, and some of that thinking is "should I look at the other three vendors I shortlisted?".
The fix is not a night shift. The fix is to answer the questions that arrive at night with the answers that already exist on your site. Most out-of-hours enquiries are not hard. They are pricing questions, integration questions, security questions and "do you support X" questions. The answer to each one is already written on a page of your own, which is what a page that answers your FAQ puts in front of the buyer at 9pm.
3. How many enquiries go cold after one unanswered question
The third number is harder to measure and the most valuable. It is the share of enquiries that received one reply, asked one follow-up, and got no answer.
Walk the 200 enquiries again. For each one, count the number of replies on both sides. An enquiry is "one and done" if it has exactly one human reply followed by silence from the buyer, or one buyer message with no human reply and no further buyer activity. Tag each one with whether the buyer's last message was a question. The ones where the last message was a question are the ones that mattered and were dropped.
This is the silent leak. The buyer asked a real question, the team was on a call or in a meeting, the day ended, and the next morning the inbox had 12 new things on top. The buyer is not sitting there waiting. The buyer is writing to the next vendor on the list. By Friday the original enquiry is below the fold and forgotten.
The fix is a queue you can see, with a timeout. If a question has not been answered in 90 minutes during working hours, it goes red. If it has not been answered overnight, it goes red at 9am. The mechanism does not need to be clever. It needs to be impossible to ignore.
4. What each of those minutes of silence costs you
The fourth number is the hardest to measure honestly, so the rule is to write the assumptions in the open and let the reader change them. Here is the working for a small B2B team selling a product with a $500 average first-year contract.
The reader's own numbers will be different. The point is the structure: every percentage point of enquiries lost to slow replies, to overnight silence, or to a missed follow-up has a dollar value, and that value is the dollar value of one enquiry times the number of enquiries lost. Plug in your own close rate and your own contract value and the arithmetic gives you the figure you can put in front of the team.
These are the reader's numbers, written as the reader's numbers, not as a market claim. The same shape works at any volume and any contract value. The point is that the three failure modes above are not equal. Overnight silence is usually the largest, because nobody sees it. Missed follow-ups are usually the second largest, because they require a process rather than a person. Slow replies during the day are usually the smallest, because the team at least knows they exist.
What the visitor actually asked on the page
The cheapest change a B2B team can make this month is to answer the question the visitor already asked before they had to ask it. Most B2B sites have a pricing page, an integrations page, a security page and a contact page, and the most common question sent through the contact form is one of those four things restated in the buyer's own words.
The way to find your own list is brutal and quick. Pull the last 100 enquiries that came in through your website. For each one, write down the first question the buyer asked, in their words. Sort them by frequency. The top five questions are the ones your site is failing to answer, because if the site answered them the buyers would not have written in.
Once you have the five, do not redesign the site. Add the answer to the page the buyer was on when they wrote. If they were on the pricing page and asked about annual versus monthly billing, the pricing page needs a sentence that says "annual billing is two months free, monthly billing is at the published rate". If they were on the integrations page and asked whether you support a specific tool, the integrations page needs the tool's logo and a one-line description of what the integration does. None of this is a feature. It is removing a reason to leave the page.
This is also the work that an AI live chat agent can absorb cleanly. A chat agent on the site answers the question the visitor has already typed, on the page they are already reading, without queuing for a human reply. The human replies are then reserved for the questions that need a human: the awkward integration question, the procurement call, the security review, the bespoke scope.
What Asyntai costs for this work
The arithmetic below uses Asyntai's published plan figures only. No other vendor appears, and no market average is invented.
Asyntai's plans are priced by the number of messages the chat agent handles in a month. The four self-serve tiers are Free at 50 messages, Starter at $39 for 800 messages, Standard at $139 for 15,000 messages, and Pro at $449 for 50,000 messages. Enterprise is a signed order form with a negotiated message allowance.
For a small B2B team, the first question is which tier covers the inbound enquiry volume the site actually receives. Most small B2B sites sit inside Starter for a useful test, then move up once they see how much the agent is handling.
The cost per message falls as volume rises, which is the usual shape for a usage-priced service. The reader's job is to pick the tier that covers the actual volume without paying for messages that will not arrive. Starter at 800 messages covers about 27 messages a day across a 30-day month, which is a comfortable fit for a small B2B site that has not run a chat agent before. Standard covers 500 messages a day. Pro covers roughly 1,667 a day, which suits a site with steady inbound traffic across multiple pages.
A worked example using these four numbers
Take the 200-enquiry month from earlier and ask what happens if the chat agent handles the questions that do not need a human. Assume the chat agent resolves half the inbound before a human is involved, and the human queue drops from 200 to 100.
The plan that fits 100 messages a month that reach the chat agent is Starter at 800 messages for $39. The cost is $39 a month to plug a $2,750-a-month leak in the example above. The reader's own close rate and contract value change the dollar figure, but they do not change the shape: a small monthly cost against a measurable monthly loss.
The figures also scale up cleanly. A site with 5,000 messages a month sits on Standard at $139. A site with 20,000 messages a month sits on Pro at $449. Each tier is published and visible on the pricing page, with no per-seat fee, no per-agent licence, and no separate charge for the underlying language model.
The four measurements in order
To put the four numbers on a single page so a small team can run them this month:
- Median reply time and 90th percentile reply time, measured across the last 200 enquiries, with the time stamp of the first human reply on each.
- Outside-hours share of enquiries, measured by binning the same 200 enquiries by the hour they arrived and counting the ones outside the published working window.
- One-and-done share, measured by tagging each enquiry with whether the buyer's last message was a question that received no human reply.
- Lost revenue per month, measured by multiplying the enquiries lost to the three failure modes above by the revenue value of a single inbound enquiry, written as the reader's own assumptions.
Run the four once at the start of the month. Run them again at the end. The change between the two is the only sales tip that matters: a number, with a working, against the same number a month ago.
What does not need measuring
Personalisation, follow-up cadence, qualification scripts and closing techniques all matter, but they do not move the four numbers above until the underlying failure modes are fixed. A perfectly personalised reply sent 14 hours after the enquiry arrived is still a 14-hour reply. A disciplined follow-up cadence does not save an enquiry that was lost overnight. A qualification script does not recover a buyer who asked a real question and got silence.
The order of work is the order of the four measurements. Fix the reply time first. Fix the overnight queue second. Fix the missed follow-ups third. Then the personalisation, the cadence and the qualification have something to land on.
The single change that costs nothing
Of everything above, the change with the lowest cost is the one at the end of the previous section: answer the question the visitor already asked on the page they were reading. It costs nothing because the answer is already written somewhere on the site. The work is moving it onto the page the buyer was on.
Do this first. It moves the median reply time without adding headcount, it answers the overnight enquiries without adding a night shift, and it removes the missed follow-ups that were really "the buyer asked a question we had already answered, but not on this page". The chat agent is the fastest way to put that answer in front of the buyer without changing the pages themselves, and the Starter tier at $39 a month covers a small B2B site's first month while you measure what shifts.
Asyntai answers the questions visitors have already typed, on the page they are already reading, with no per-seat fee and no setup project. Starter is $39 for 800 messages a month, Standard is $139 for 15,000, Pro is $449 for 50,000, and the Free plan covers 50 messages for a test run. Enterprise is a signed order form with a negotiated allowance.
Plan prices and message allowances: Asyntai pricing page, checked September 2026.
The four measurements and the worked example use the reader's own volume, close rate and contract value as inputs. Replace any figure with your own and the arithmetic follows.