Ask every enquirer where they heard about you and write the answer down. That one question closes a gap no analytics tool can.

What the reports cannot see

Analytics measures visits to a website, which is a subset of how a small business actually gets work.

A telephone call from somebody who was told about you by a neighbour appears nowhere.

Neither does a van seen on a street, a card on a noticeboard, a mention in a local group, or a customer returning after three years.

Which means a business managing by its analytics is managing the channel it can see rather than the one producing the work.

The question that closes the gap

One sentence, asked of every enquiry, and it is the whole method.

How did you hear about us, asked naturally during the first conversation rather than as a form field.

Write the answer down somewhere consistent, in the customer's own words rather than sorted into categories.

After thirty or forty answers the picture is clearer than anything a tool will produce, because it covers every channel rather than one.

Most businesses ask this occasionally, remember the interesting answers, and record none of it.

The recording is the part that matters, since memory over-weights the recent and the surprising.

Where to record it

The fourth is what turns a count into something useful, because a channel producing many enquiries and no work is a different proposition from one producing three enquiries and three jobs.

Expect the answer to be vague

Since people genuinely do not remember, and that is information too.

I found you online covers a search, a listing, a directory, a social post, and a link from somewhere else.

Somebody recommended you is more useful and worth one follow-up question, since knowing who refers you is worth more than knowing the channel.

Where the answer is genuinely I cannot remember, record that rather than guessing, since a large proportion of those is itself a finding.

Do not push, because this is a conversation with a customer rather than a survey.

One question and whatever they say is the right level of effort.

A worked example

A business had good analytics and could account for about a third of its enquiries.

They added one column to their enquiry list and asked everybody for three months.

Of about seventy enquiries, roughly a quarter said a specific person had recommended them, and eleven of those named the same three customers.

Another group had seen the van, which nobody had considered a channel at all.

The website accounted for less than analytics had implied, because a proportion of people search for the business by name after hearing about it elsewhere.

They spent the following year's budget differently, and thanked the three customers properly.

Searching your name is not a source

The distinction that changes how the whole picture reads.

Somebody who types your business name into a search engine already knew about you, so the search was the route rather than the cause.

Analytics counts that as a search visit, which inflates what the website appears to be producing.

The cause was whatever told them your name: a recommendation, a van, a sign, a previous job.

Separating branded searches from everything else in your reporting takes a few minutes and makes the figures considerably more honest.

It also usually shows that the site is doing less of the finding and more of the confirming, which changes what to improve.

Track the calls at least roughly

Since the telephone is the largest gap for most trades.

A separate number on the website, forwarded to your usual line, tells you how many calls the site produced without any other setup.

That is a small monthly cost and it closes the largest measurement gap available.

Where that is more than you want, simply asking every caller how they found you and noting it does most of the same job.

And check whether your business listing reports calls, since many do and almost nobody looks.

Any of those is better than the current position, which is a number nobody has.

Read it once a quarter

Rather than continuously, since the value is in the pattern.

Count the answers by rough category, look at which produced work rather than only enquiries, and note anything that changed.

Compare against the previous quarter, since the useful signal is movement rather than the absolute split.

Then decide one thing on the strength of it, which is usually where to spend or where to stop.

Twenty minutes, four times a year, against a column somebody filled in as they went.

Thank the people who send you work

The obvious use of this data and the one that gets skipped.

Once you know who refers you, and it is usually a small number of people, that is a list worth acting on rather than filing.

A short note saying you have noticed and are grateful costs nothing and is remembered, particularly since almost nobody does it.

Do not attach an incentive, since paying for referrals changes the relationship and frequently reduces them.

Simply knowing which three customers or which neighbouring trade account for a quarter of your work is worth the whole exercise.

The counter-case

This data is soft.

People misremember, give the answer they think you want, and compress several touches into whichever they recall last.

It is also not comparable to analytics figures, so combining the two into one chart produces something misleading.

And for a business whose work genuinely all comes through the website, the gap being described does not exist.

Ask the question, record the words, separate branded searches, and read it quarterly.

Starting in January

  1. Add one column to your enquiry list.
  2. Ask everybody, in conversation.
  3. Record their words, not a category.
  4. Note whether it became work.
  5. Separate branded searches in reporting.
  6. Check whether your listing reports calls.
  7. Read it quarterly and decide one thing.

Step five changes what the existing figures mean, since somebody searching your name already knew about you and the site was confirming rather than finding.

What can honestly be established from a small number of enquiries is covered in attribution for a business with one salesperson.


Frequently asked questions

What can analytics not see?

Everything that is not a website visit: a call from somebody told by a neighbour, a van seen on a street, a card on a noticeboard, or a customer returning after years.

What is the method?

Ask every enquirer how they heard about you, in conversation, and write the answer down in their own words. Thirty or forty answers beats anything a tool produces.

What should I record?

One column in whatever list you keep: their words, the date, and whether it became work. That last one distinguishes many enquiries from actual jobs.

What if the answer is vague?

Record it as given. A large proportion of I cannot remember is itself a finding, and somebody recommended you is worth one follow-up about who.

Why separate branded searches?

Because somebody typing your business name already knew about you, so the search was the route rather than the cause. It inflates what the website appears to produce.

How do I track calls?

A separate number on the website forwarded to your usual line, or simply asking every caller. Also check whether your business listing reports calls, since many do.

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.

Can you account for where your work comes from?

Add one column to your enquiry list in January and ask everybody. Thirty answers will surprise you.

Start a Conversation