Leads

Ask a small business how many enquiries it received last quarter and you will usually get a number that is a feeling. Ask which advertising produced them and you will get the one someone remembers. Ask what proportion converted and there is frequently no answer at all.

This is not carelessness. It is what happens when enquiries arrive through four channels into an inbox, a telephone and a form, and the only place they are all recorded is in the heads of whoever handled them. There is no ledger, so there is nothing to count.

The cost is not that the questions go unanswered. It is that decisions get made anyway, on impressions. Advertising budget goes to the channel that produced the enquiry somebody remembers most vividly, which is usually the most recent or the most expensive. Staff time goes to the enquiries that feel urgent instead of the ones that convert. And no one notices when a source that used to work quietly stops.

The three things worth recording

You do not need a large amount of information to fix this. You need three things recorded consistently, and the discipline to record them when the enquiry arrives rather than later.

Where it came from. Recorded at the point of capture, from a fixed list rather than free text, because a source field where people type whatever they like produces eleven spellings of the same thing and cannot be counted.

Who owns it. One person, named. The most common way an enquiry dies is not that two people chased it, it is that both assumed the other had, and an unowned lead is one nobody is going to ring.

What happened. Converted, or not, with a reason. That last part is the one everybody skips and the one that pays, because it is the only route to knowing whether you lose on price, on speed, or on something you could fix this month.

Enquiries by source is the least useful of the three

It is the number everybody starts with and the one that misleads most often, because the sources that produce the most enquiries are frequently the ones that produce the worst. A cheap listing site can generate a great deal of volume and almost no work, and if you measure volume you will spend more on it.

Conversions by source is the number to spend against

The same report, filtered to the enquiries that became customers. This is the one that changes decisions, and it frequently reverses them. A source producing four enquiries a month of which three convert is worth considerably more than one producing thirty of which two do, and the second one usually costs more.

Stage history tells you where the process is broken

How long enquiries sit at each stage before moving. This is the diagnostic, and it is almost always surprising. Most organisations believe they lose enquiries at price. The stage history frequently says they lose them between first contact and the second conversation, which is a responsiveness problem, costs nothing to fix, and would never have been found by asking people what they thought.

The reason field, read as prose

Reasons on lost enquiries are the least numerical and most valuable output. Read twenty of them in a row and patterns appear that no report would produce: a competitor mentioned repeatedly, a service you do not offer being asked for weekly, a price point that keeps appearing.

This is also why lost enquiries are kept rather than deleted. Beyond the reporting, the people who enquired and did not buy are the warmest list any organisation owns, and because consent was recorded with a date and a source, contacting them next year is something they agreed to instead of something you hope is acceptable. Whether you may write to them at all is settled by the Privacy and Electronic Communications Regulations rather than by a CRM. What the product does is hold the evidence.

Record everything, including the obvious rejections

The instinct is to record only the enquiries that might go somewhere, which destroys the denominator. Twenty seconds on an enquiry that is obviously not for you, marked lost with a reason, is what tells you six months later that a fifth of your enquiry volume is unsuitable and where it comes from.

Fix the source list before you start

A short list of sources that people choose from, agreed in advance. Not free text. Six or eight options covering how work actually reaches you, with one called Other that somebody reviews occasionally. A source field where people type freely produces Google, google, Web, Website and Search, and none of them can be counted together.

Give every lead an owner immediately

Even if the owner is provisional. An enquiry with no owner is an enquiry no one will ring, and it is the single most common way work is lost in a small organisation. The allocation being wrong is a much smaller problem than the allocation being absent.

Build one view and look at it on a Friday

Enquiries at any stage with no activity for ten days. That is the list. It takes two minutes to build and it is the highest return thing anybody does with this part of the product, because those people have not decided yet and nobody has spoken to them.

No lead scoring. Nothing assigns a number to how promising an enquiry is. A score nobody can explain is a score no one acts on, and a score anybody can explain is a saved view with fewer moving parts.

No automatic routing rules. No assignment by territory, by round robin, or by product. At this size an owner is chosen by the person taking the enquiry, who usually knows something a rule does not.

No web tracking. Nothing follows somebody around a website, records which pages they read, or attributes an enquiry to a sequence of visits. That is a different category of product with different consent implications and it sits badly beside recording consent as a fact with a source.

No contact database or address finding. No buying a list, no enrichment, no automated prospecting. This is a position instead of a gap.

No merging. Duplicates are found at capture and shown to you, and nothing is ever joined up for you. If the same person enquires twice you still have two enquiries, both pointing at the same record. That is deliberate: merging is a judgement about your own business and a destructive operation, and the two enquiries are two real events even when they came from one person.

The enquiry is from someone already on your books

At the moment of conversion the email address is checked against the relationships you already have. If it matches one, the enquiry joins that record instead of creating a second version of the same customer. This is the failure people notice, because their history sits under one record and their new work under the other, and once anything has been filed against both it cannot be undone by hand.

The same person enquires three times in a year

You get three enquiries, and all three lead to the same relationship. What first brought them in stays as it was; what they most recently came through is updated. Both are kept deliberately, because they answer different questions. One says what got them into the room and the other says what made them act this time, and an organisation choosing which of those to spend against needs both numbers instead of an average of them.

The enquiry is a company instead of a person

If a company name was recorded on the enquiry, conversion creates an organisation and names it after the company. If not, it creates a person. That is decided by what somebody recorded rather than by inspecting the name for something that looks like a company, which is a guess that is wrong about sole traders roughly whenever it matters.

Somebody deleted an enquiry and now wants it back

Deleted enquiries go to a recycle bin and wait thirty days before they are removed for good. Anyone who may work an enquiry may bin it and may restore it, because a bin is reversible. Restoring puts the record back into search as well as back into the list.

Whilst an enquiry is in the bin it cannot be qualified and it cannot be converted. It is treated as absent rather than quietly resurrected into a relationship, which is the behaviour you want the first time someone bins a spam enquiry and someone else clicks convert on a stale screen.

Somebody tries to convert the same enquiry twice

It is refused. An enquiry that has already converted points at the relationship it produced, and converting it again would create the second record the email check exists to prevent.

A stranger ticks the consent box using somebody else's address

The address on a public form is typed by whoever is filling it in and nothing verifies it. So a form submission never writes a consent answer against a record that already exists. It only records consent for a record the submission itself created.

The reason is the version of this that does damage: somebody who knew a customer's email address could otherwise submit the form with the box left clear and quietly remove a real customer from every future campaign. The enquiry is still created, and it still shows what was ticked, so a member of staff can see it and act on it. The proper answer is a confirmation message before any consent is recorded at all. That is worth building and it is not what happens today.

An enquiry is its own record, not a flag on a relationship

The obvious design is one table of people with a field saying whether they are a lead or a customer. It was rejected. Most enquiries never become anything, so that design fills the list of people you have relationships with mostly with people you do not, and every count, every export and every audience then has to remember to exclude them.

The cost of the decision is real and you will feel it: there are two places a name can be, and someone looking for a person has to accept that search returns both kinds. We took that over a relationship list that is four fifths strangers.

Conversion is all of it or none of it

The alternative is to create the relationship immediately and let the piece of work be added afterwards, which is simpler and is what happens when the two steps are separate screens. It was rejected because the failure is invisible. A converted enquiry pointing at nothing looks exactly like a converted enquiry, and the way you find out is a customer ringing about work nobody has a record of.

So if any part of the conversion cannot complete, none of it does. The visible cost is that a practice with no pipeline configured gets a refusal at the moment it least wants one, rather than a partial success it would not notice.

Where someone came from is copied, not looked up

The tidy version stores the source once, on the enquiry, and joins back to it from the relationship and the work. It was rejected because nothing downstream ever looks back at the enquiry. By the time anyone asks which source produced a piece of business, they are looking at the business, and if the answer is not there it is not an answer.

Copying data is normally the wrong instinct and here it is the right one, because the fact being copied does not change. Where someone originally came from is fixed the moment it happens.

Two answers about origin are kept, not one

Most products record one attribution field and it quietly answers whichever question the person who wrote it had in mind. There are two questions. What first brought this person into contact with you, and what they came through this time, are different, and the gap between them is frequently a year and two different budgets.

Both are kept on the enquiry, and both survive conversion onto the relationship and onto the work. The uncomfortable half is that you now have two numbers where you expected one, and the first meeting where they disagree is an argument. That argument is the useful part.