A two-week audit: 90 days of CRM data, 3 team interviews, call analysis, a review of proposals and the client base. Below is the real deliverable, with the company name and people's names taken out.
The four numbers above are read from the CRM and the interviews. They are not calculated and not interpreted. The scale has to be said up front too: six deals in 90 days means every percentage in this audit sits on very small numbers, and is therefore fragile.
The left column is read from the CRM and the interviews. The right column is the target state, not a promise - every line has a measurement that will say after 90 days whether it happened.
Three steps that are the same in every project, whatever the findings are.
We go through the findings with the owner, confirm priorities, make the first week's decisions, and name who owns what.
One session · First weekFixed 30 min sessions with each team member once a week. Numbers, open proposals, support. The owner gradually takes these over.
3 sessions a week · 90 minEvery action is marked not started, in progress or done. Updated every week. An Excel tracker for the client to review.
Throughout the projectActions that free up time and create a baseline. None of them depend on individual skill development, and none of them is revenue on its own.
An interview says that preparing proposals takes 8-16 hours a week. But the CRM shows 12 proposals in 90 days, which is about one a week. At 2-4 hours per proposal that is 2-4 hours a week, not 8-16. Over a year that is roughly 97-195 hours in total, and the saving from 3 hours down to 30 minutes is about 120 hours a year, not 440.
Track the actual time on every proposal for two weeks and check whether the hours go into proposals or into other work that was counted together in the interview. After that the capacity number gets rewritten against the fact. Until then this audit uses the more conservative version, 120 hours a year.
There are 6 closed deals in 90 days, 4 of which came through referrals. That is a fact, and it is an important signal. But "60% of the business comes through referrals" is a percentage of six deals, and one deal either way moves it by 17 percentage points. You cannot build precise models on that.
Widen the data window to 12 months before any referral model is treated as credible. If a year of data shows the same ratio, the signal becomes a fact. If not, the referral line gets recalculated.
The CRM was rolled out recently. An interview says plainly that there are no call stats and that deal information comes from the bookkeeper at the start of the month. In parallel the team works in Excel, WhatsApp and from memory. That means the baseline we will measure against in 90 days does not exist yet.
The baseline is recorded in the first two weeks: number of proposals, number of follow-up contacts, number of calls, closes. Without it, every later comparison is against memory, not against data.
Every proposal is built by hand, copying insurer quotes into a presentation. 2-4 hours per proposal, 12 proposals in 90 days. The average time from meeting to proposal is 12 days, and that is directly tied to this work.
Build the process, a template for the output, a demo and tuning. The measurement is actual time per proposal before and after, plus the average time from meeting to proposal. Where the freed-up time goes is decided separately and written down.
Freed-up hours are not revenue. They turn into money only if they are redirected into a specific sales activity that converts. This audit does not model them as revenue, because there is no data on what one freed-up hour is worth in this company. The real gain that can be promised is a faster proposal for the client, not higher revenue.
The setter calls a colleague to find out which times are free before anything can be agreed with the client. That is two steps where there could be one, and the time does not get locked in during the call, while the client is still on the line.
A shared calendar with a booking link and automatic reminders. Measurement: how many meetings get booked during the call, and the no-show rate before and after.
The CRM was rolled out recently, and in parallel the team works in its own system: Excel, WhatsApp and from memory. Without entry standards the weekly sessions will not work, because there will be nothing to look at.
A standards document and a weekly checklist, one page. The owner verifies it in the weekly review. The measurement is how complete the entries are, not a subjective opinion.
In the interview the person says himself that his main job is a support function, not sales. Zero new deals in the new segment in 90 days, and just under 2% of the client base called. The central comment: no time, because it goes on small requests from existing clients. In practice he runs a split role between a support function and his own client portfolio, and that works.
A written role document, a daily structure with focus blocks, and a task flow protocol from the owner. Cold calls into the new segment come out of the role, because 90 days produced no result. Measurement: task turnaround time and the size of the personal portfolio.
Two systems that use what the company already has: 500 existing clients and proposals already sent. These are also the only two lines in this audit with a money figure.
Of the 6 deals in 90 days, 4 came through referrals, with no system behind it. There are several hundred clients in the base, and not one of them has ever been asked for a referral. If the channel that brings in most of the business runs on chance, fixing it is the most logical first step.
A script, a process document and a separate cycle in the CRM. 30 days after a deal closes: satisfaction, a review, the referral ask. In the first 30 days the goal is not revenue but baseline data from 50 calls into the existing base: how many people answer the question at all, and how many give a contact.
60% is a percentage of six deals. One deal either way moves it by 17 percentage points, so this is a signal, not a measured fact. The 10-15 referrals a year and the 40% close rate used in the model do not come from this company's history, because there is no such history.
11 of 12 proposals in the last 90 days sit with no follow-up. There is no definition of when and how to go back to the client after a proposal goes out, so every time it is a separate decision, and usually it does not get made.
A protocol and email templates with a 3, 7, 14 and 30 day cadence, plus call scripts and training. Execution is measurable straight away: how many proposals got the full cadence. The result is only measurable after 90 days, and at 12 proposals a quarter it will take several quarters to become credible.
The team has no weekly or monthly targets covering activity, process or result. Business results only become visible at the start of the month, from the bookkeeper. That means there is up to four weeks between an action and the feedback on it.
Three metrics for every role: activity, process, result. A clear agreement on how they get acted on. The owner approves them and communicates them. Measurement: whether the metrics are logged every week without a gap.
B2B clients who hold the core product are not systematically offered other products. Same story in the private client portfolio. Today additional products get offered now and then.
Segment clients by product in the CRM, write scripts for 1-2 products, train on them. One campaign into one segment in the first quarter, to get a real response rate to build a model on later.
This line has no money figure. To model it you would need to know how many of these clients already hold other products elsewhere, what the average commission is by product, and how much time a campaign takes. None of that exists.
Call analysis shows that cold calls lack the structure to hold the conversation after the first objection, and that calls move into a pitch at around the first minute. On the cold calling side the current result is roughly 250 calls per booked meeting and one new deal in 90 days.
30 days of self-directed drilling: one specific question, 15-20 minutes a day, used in at least 80% of calls, a weekly report. In parallel, call logging starts in the CRM so the next cycle has a real baseline. If level 1 gets done, level 2 follows. If not, the owner gets clarity on how to steer the team's development.
The call sample is small, and 250 calls per meeting is an estimate from an interview, not from the CRM, because there are no call stats. A skills gap is a plausible explanation, but so are list quality, segment choice or a mismatched offer. For that reason the skills line has no money figure.
Two observations that need a decision from the owner. The audit prepares the analysis and the options, the choice stays with the company. Neither has a money figure, and both rest on weak evidence.
In the interviews, team members said that urgent tasks arrive without clear priorities, which makes it harder to stay focused on the core role. That is the subjective view of three people, not a measured process.
A workflow document, a task management tool and a daily rhythm protocol. Measurement: how many tasks come in outside the defined flow each week. This is a habit change that takes time, and within this audit it does not count as a quick win.
The variable part of pay is small relative to base pay for all three team members. An internal comparison shows that the person with the largest commission share in their own portfolio is the more active one. That is one observation in one company, not a proven relationship. The actual pay figures are not shown in this public version.
A recommendations document with an analysis of the current structure, the internal comparison and 2-3 alternatives with simulations. The decision sits with the business owner, and the recommendation is a starting point for discussion, not a finished change.
The first version of the audit claimed here that the variable part is lower than the industry norm. No publicly available, comparable pay data for Latvian insurance brokers was found, so that claim has been removed. Without a source it is an opinion, not a finding.
Two lines with a money figure and two deliberately without one. Each one shows its formula, the assumption behind it, and the test that would cancel it. All numbers are in commission revenue.
10-15 referrals a year × 40% close rate × roughly €1,800 average commission = €7,200 to €10,800
Both headline numbers are assumptions. Neither the volume of 10-15 referrals from a base of several hundred clients nor the 40% close rate comes from this company's history, because that history was never recorded. The supporting signal is 4 of 6 deals through referrals, but that is a sample of six deals.
If the first sample of 50 calls into the existing base produces fewer than 2 referral contacts, the volume assumption is wrong and the line gets recalculated downward, or to zero.
Roughly 49 proposals a year, of which the cadence saves 1-3 extra deals × €1,800 = €1,800 to €5,400
Assumes proposal volume stays similar and that follow-up affects the outcome at all. At 12 proposals a quarter, a difference of one or two deals is inside statistical noise, and only several quarters will give a credible answer.
If after two quarters with the full cadence the number of closes has not changed, the line is zero. At this volume you cannot tell any sooner.
Roughly 49 proposals a year × a 2.5 hour saving = about 120 hours. The first version of the audit stated 440 hours, which does not match the number of proposals, so it has been corrected.
Hours are not revenue. They turn into money only if they are redirected into a specific activity that converts. What one freed-up hour is worth in this company is not known, so there is no money figure.
If two weeks of time tracking show that the actual time per proposal is closer to 1 hour, the saving drops by roughly half.
There is no calculation. It is not known how many of these B2B clients already hold other products elsewhere, or what the average commission is by product, and call stats are not logged.
Any number here would be made up. Baseline data gets collected in the first quarter, the model comes after that.
Nothing to cancel, because nothing has been claimed.
The modeled numbers are not a forecast and not a guarantee. They are in commission revenue, not profit, and the engagement fee has not been deducted. The sample in this audit is especially small, six closed deals and twelve proposals in 90 days, so every percentage is fragile and the ranges are wide. A credible answer on the revenue effect comes from a 12-month comparison, not a 90-day one. If an assumption does not hold up, the number gets recalculated, not defended.
Two weeks, fixed price. CRM data, team interviews, call analysis, and a model with every assumption and every test that would cancel it set out next to it.