Sales Process Improvement That Starts With One Leaky Stage

What is improving a sales process, in plain terms?

Sales process improvement means finding the one stage of your pipeline that loses the most people. You change one thing at that stage and measure it again. You keep the change only if that stage's number moves. Then you write the change down so every rep runs it the same way. Tools and training come after you know where the loss is.

Most companies believe they already have a process. A 2025 Bain survey covered over 1,200 senior commercial leaders. More than 80% said their sales and marketing work follows a repeatable structure. Only about 20% had got the full value from it (Bain & Company). Writing the steps down is easy. The work is knowing which step leaks.

My team books meetings for B2B clients, mainly via cold email and LinkedIn, and we never cold call. The meetings below were booked through email and LinkedIn, and the client's own team ran them.

How do you find where your sales process leaks?

Count every stage for one fixed period (if you have no stage definitions yet, start with what a sales process is). Divide each stage by the stage before it, then look for the ratio that falls hardest. Here are two real 30-day funnels from our campaigns.

I computed every rate from the raw counts, rounded to one decimal.

The leak is the stage with the lowest ratio. In both funnels that is reply to interested, at 7.3% and 8.4%. Every stage after it converts at 18% or better.

Some stages always convert lower than others, so also compare each stage with the same stage in your previous month. With 3 and 2 closed deals the sample is small. Still, neither funnel points at the closer first. Both point at getting more repliers to want a conversation.

A channel split shows the same thing from another angle. Here is one month, April 17 to May 17, 2023, for a UK firm that sells software development and IT consulting, email and LinkedIn rows only.

After a positive reply, the two channels behave almost the same, so the gap in total sales comes from how many positive replies each channel produces.

Is the problem before the meeting or on the call?

Booking a meeting and winning the deal are separate jobs with separate owners. Everything up to "showed" belongs to outreach: list, sender, message, follow-ups, reply handling and confirmation. Once the prospect joins the call, it belongs to your sales team.

Make that split before anyone argues about blame. Booking is not where these two funnels lose people: they book 76.2% and 53.8% of the leads who showed interest. Measured against all replies, only 5.5% and 4.5% booked, because most replies never turn into interest.

The show stage sits between the two owners. Neither funnel above lost much there. If yours does, treat it as its own leak; the fix is in the table below.

Decision rule: if the biggest drop comes at or before "showed", fix outreach. If people show up and do not sign, fix the call and leave outreach alone until the close holds.

Why won't more leads fix a weak stage?

A funnel that fails at 1,000 contacts also fails at 10,000. Volume multiplies the ratios you already have, broken ones included. A weak closing ratio gets weaker as more calls land on the same team, which is why more booked calls make a weak close worse. Paying for a hundred extra meetings to dodge a sales problem only makes it bigger.

You look at the 2 or 3 contracts that got signed. I look at everyone who slipped away, because that is where the next deals hide. In Funnel 1, 9 people sat through a whole call and did not buy. And 268 replies never turned into interest.

Fix the funnel first. Then scale. I raise volume only after the weakest stage has held its new rate at the current volume.

Which fix matches which leak?

When the top row is the problem, I never start with copy: good copy cannot save a poor list or an unreliable sender. For the booking row, my team puts one person on reading and answering replies, so no interested lead sits waiting.

The last row is easy to skip. My team sees the third follow-up do the most work, provided the gaps between messages are real ones.

How do you test a fix without fooling yourself?

Change one variable per window. Keep volume, list source and sender the same, and compare the same stage ratio with the previous window. Change two things at once and you will not know which one moved the number.

A worked example, with Funnel 1 as the baseline: reply to interested sits at 7.3%. For the next 30 days, change only the offer in the first message, and keep about 1,200 contacts from the same list source and account.

Then divide interested by replies again. If the ratio climbs to around 10% on a similar number of replies (a hypothetical target), do not keep the change yet.

Run the window once more first. With about 289 replies, a move from 21 to 29 interested can be noise. Kissmetrics' guide to A/B test sample size says that at the standard 5% significance level, about one test in twenty can show a false positive even when nothing real has changed.

Small changes need more observations to confirm. If the ratio stays near 7%, test the next single variable.

Log this in your CRM for every window:

• Start and end dates
• The one thing you changed, in one line
• Counts for contacted, replied, interested, booked, showed and closed
• The ratio you are testing, before and after

I tell clients to use a CRM so no lead slips away; which one matters far less than using it. Keep the logging light. Forrester's sales activity study found that administrative tasks take up over a quarter of a rep's week (Forrester).

Log counts per stage, never essays. No system yet? Read why the CRM comes before the first campaign.

How do you make the improved process repeatable?

Consistency comes from writing down the change that won. Put it in two places: the playbook every rep reads, and the stage definition in your CRM, so that "interested" means the same thing whoever logs it. The same Bain survey found that 70% of companies have not built their sales plays properly into their CRM and revenue tools.

A fix that lives in one rep's head leaves with that rep.

Then write the process for the buyers you actually have. In CEB research published in 2017, the average buying group for a B2B solution grew from 5.4 to 6.8 people (Harvard Business Review). "Showed" may mean one person out of six or seven, so write down who joins the second call and what each of them receives.

Keep the order fixed: conversations first, then calls, then clients. Once the outreach stages hold, judge the close row against meeting-to-close benchmarks.

This week, put last month's counts into the six rows of the first table. Find your lowest ratio and choose the one change you will test there.

If you have a month of counts per stage and cannot tell where the leak is, book a call and send them over. I will tell you where I would look first.