A diagnostic, in full.

Six interviews of forty minutes each, held over four working days.

• ·Rachel Okonjo, Managing Director
• ·Tom Barr, Sales Director
• ·Two estimators, interviewed separately
• ·The office administrator who opens the shared inbox each morning
• ·The contracts manager who inherits a job after it is won

We also examined ninety days of enquiry records, the CRM as it is actually configured, the shared enquiries inbox, the enquiry routes on the website, and twelve recent quotations chosen at random across won, lost and unanswered.

We did not examine pricing strategy, margin on delivered work, the tender portal subscriptions as a commercial decision, or anything after a job is won. Those were out of scope, and nothing here should be read as a view on them.

Kestrel does not have a lead generation problem. It receives more qualified enquiry than it converts, and the loss happens in a twelve day window between an enquiry arriving and a quotation reaching the customer.

The largest single improvement available costs nothing and needs no software. We recommend one narrow purchase, described in section 6, and we recommend against the two purchases Kestrel was already considering.

The written process says every enquiry is logged in the CRM on arrival, assigned to an estimator, and quoted within five working days. The observed process differs in four ways.

Four inbound routes, two of them unmonitored

The website form and the shared inbox both reach the administrator and both are logged. The main phone number is logged when whoever answered remembers to. The trade portal notifies one estimator’s named work address, because that is how the account was set up in 2023, and those enquiries never enter the CRM. They are worked, and several are won, but they are invisible to reporting. Across ninety days we could account for 214 enquiries in the CRM and identify at least 61 more that existed elsewhere.

A private spreadsheet

One estimator has kept his own tracking spreadsheet since March. He started it because the CRM requires a project value before a record can be saved, and at first contact he usually does not have one. His workaround is sensible and he was not hiding it. The consequence is that the CRM understates live enquiry by roughly a third, and the Monday pipeline report is drawn from the CRM. Every number Kestrel uses to make decisions has a denominator missing a third of its enquiries.

No owner between acknowledgement and quotation

The administrator sends the first reply within a few hours and does it well. After that the enquiry sits in a shared state. Both estimators described working the enquiries that chase them. Neither could name who is accountable for an enquiry that goes quiet, and neither thought it was them.

Quotation speed is set by estimator load, not by value

Quotes are worked in roughly the order they are picked up. A £4,000 flooring replacement and a £110,000 fit-out receive the same queue position.

The arithmetic below uses Kestrel’s own ninety day records. Every assumption is named so it can be argued with. If an assumption is wrong, the conclusion should be recalculated rather than discarded.

Enquiries that are never quoted

76 in the quarter. If a quarter of those were viable and converted at the slow rate of 14%, that is roughly 2.7 jobs a quarter, or about £50,000 of revenue.

The one in four viability figure is our estimate from reading the enquiry text, and it is the weakest number in this brief. Kestrel can replace it with a real one by reviewing thirty unquoted enquiries. Do that before acting on the size of this number, though not before acting on the cause.

Quotation speed

If the quotes currently issued after eight days went out within three, and converted at the fast rate rather than the slow one, the difference across 71 quotes is about 12 additional jobs a quarter.

That figure is almost certainly too generous. Fast quotes are probably faster partly because they are simpler, so some of the gap is the job rather than the speed. Halve it and it is still six jobs a quarter at a median of £18,400.

Neither number should be given to a supplier as a savings target. They are here to establish that the twelve day window is where the money is, which is the decision this brief is for.

In this order. The first two are the ones that matter.

1. 01

   Repoint the trade portal notifications

   Change the notification address on the portal account to the shared inbox. Ten minutes of one person’s time, and it closes the largest visibility gap in the business. Do this first, because every measurement below is wrong until it is done.

2. 02

   Name an owner for the middle of the enquiry

   One person, by name, accountable for every enquiry between the acknowledgement and the quotation. Not a rota. At Kestrel’s size the administrator is the right choice, because the work is chasing and scheduling rather than estimating. Give her the authority to ask an estimator for a date and to escalate when she does not get one.

3. 03

   Make the project value field optional in the CRM

   A configuration change, not a project. It removes the reason the private spreadsheet exists. Then agree that the spreadsheet stops, and check in a fortnight that it has, because it will not stop on its own.

4. 04

   Set a triage rule for quotation order

   Anything above a threshold Kestrel sets, we suggest £25,000, is quoted first. Everything else is worked in arrival order. One line, no software.

5. 05

   Start measuring one number

   The percentage of enquiries that receive a second contact. Count it by hand each Friday for six weeks. It is currently around 20%, and nobody in the business knew that before this diagnosis.

Expect the first four to take under a day of combined effort. If they take longer than a week, something is being turned into a project that should not be.

Mostly nothing.

Section 5 addresses the causes. Buy nothing until those have run for six weeks and the second contact number has been measured, because buying now would attribute the improvement to the purchase.

One purchase is worth making after that, and only if the measurement supports it. Kestrel needs enquiries from all four routes to enter one system automatically, with an ownership state and an age. That is a small integration and a CRM configuration job. It is not a new CRM.

Insist on this in writing before signing

1. 01The supplier configures the existing CRM. Any proposal that begins with a platform migration is answering a different question and should be declined.
2. 02A fixed price for a defined scope, with the scope written as the four inbound routes by name. Day rates on a job this size transfer all the risk to Kestrel.
3. 03The trade portal integration is demonstrated working on Kestrel’s own account before final payment, not described in a specification.
4. 04Kestrel owns the CRM account, the integration credentials and any automation configuration, confirmed in the contract. Suppliers who hold the account are expensive to leave.
5. 05A named person and a completion date. On a job this size, four to six weeks.
6. 06No ongoing licence for the integration layer beyond the tools Kestrel already pays for.

A realistic price for that scope is £3,000 to £7,000. A quote above £15,000 is quoting a different project, and should be sent back with this brief attached.

Do not buy a new CRM

Both firms Kestrel spoke to proposed one. The current CRM is not the problem. A migration would consume two quarters and move the same broken process into a better looking system.

Do not buy a lead generation service

Kestrel is already failing to quote 76 enquiries a quarter. More enquiries would increase the loss.

Do not automate the follow-up before the owner exists

An automated second email sent by nobody and replied to by nobody is worse than the current state, because it looks like contact in the reporting.

Baseline these now, before anything changes.

If the second contact figure has not moved by week six, the owner named in section 5 does not have the authority the role needs, and that is a management problem rather than a systems one.

Four things would change the conclusions, and we could not verify them.

• ·The one in four viability estimate for unquoted enquiries is our reading of the enquiry text, not Kestrel’s data. Section 5 says how to replace it with a real number.
• ·The win rate difference between fast and slow quotes is correlation. We have not separated job complexity from response speed, and some of the gap is certainly complexity.
• ·We interviewed two estimators. A third works mainly on site and we did not reach him. If he holds a fifth inbound route, section 3 undercounts.
• ·Ninety days spans one season. Kestrel described autumn as its busiest period, so the quotation delay may be at its worst in this sample.

None of these change the recommendation in section 5, which costs nothing and is worth doing under every version of the numbers above. They do affect the size of the prize, so treat section 4 as an order of magnitude rather than a forecast.