A useful “How did you hear about us?” question gives people a short list of places they recognise. Start with search, social media, recommendations, and the channels you use. Include “Other” and “I don't remember.” Save the answer with the enquiry, alongside any source data your website captures.
TL;DR
Ask about one moment. First discovery, today's visit, and the reason for buying are different questions.
Make the choices easy to recognise. Use a short list, avoid overlapping options, and let people say they don't remember.
Keep both sources. A person's answer adds context. Recorded source data preserves details they may forget. Compare both with sales outcomes.
A short list of survey options you can use
For a contact or quote form, start with this version. Adjust the channels to fit your business.
How did you first hear about us?
Choose the closest answer. This question is optional.
- Google or another search engine
- Social media
- A recommendation from someone
- An email or newsletter
- An event
- Other — please tell us where
- I don't remember
This is a starting list, not a rule that every form needs seven choices. If podcasts or AI tools are relevant, add them. If you never use events, remove that option. Read the “Other” answers before deciding what else belongs in the list.
If someone chooses social media, an optional follow-up can ask which platform. If they choose Other, give them a short text box. Don't require a detailed explanation just to send an enquiry.
Choose options that fit your customers
Use words your customers would use. “Google” is easier to recognise than “organic acquisition.” Someone who saw a LinkedIn post should not need to know whether it was paid.
The examples below show how to adapt the starting list. Replace a broad option where useful, rather than adding every row to one form.
| Your business | Choices worth considering | A useful optional follow-up |
|---|---|---|
| Local service business | Google Search or Maps; recommendation; local directory | Which directory? |
| Business software | Search engine; LinkedIn; software directory; integration marketplace | Which directory or marketplace? |
| Consultancy | Recommendation; event; podcast; article | Which event, show, or article? |
| Business with AI referrals | ChatGPT or another AI tool | Which tool? |
Keep choices distinct. Offering both “Social media” and “Instagram” makes an Instagram visitor choose between two correct answers. Either list individual platforms or keep one social option and ask for detail afterwards.
The same problem appears with “Google” and “Online advertisement.” A Google ad fits both. Ask where someone saw you first, then collect ad details separately if they know them.
Pew Research Center's survey guidance recommends clear, specific questions and answer categories that do not overlap. It also explains that answer order can affect responses. If your form supports it, consider shuffling the channel choices while keeping Other and “I don't remember” at the end.
Ask about discovery, the visit, or the buying decision
Small wording changes can change what you're measuring. Choose the question based on the decision you want to make.
| What you want to learn | Question to ask |
|---|---|
| Where someone remembers discovering you | How did you first hear about us? |
| How they reached your site this time | How did you find our website today? |
| What made them enquire | What made you decide to contact us? |
For example, someone hears your name on a podcast. A week later, they search for it on Google. Your pricing then persuades them to request a quote. All three answers are useful, but they describe different parts of the same decision.
For a simple source question, start with first discovery. The word “first” makes the request clearer. It still asks for a memory, so it cannot prove the first time someone encountered your business.
Should people select one answer or several?
Use a single-choice question when you want the earliest source someone remembers. Radio buttons show all the choices; a dropdown takes less space.
Use checkboxes if you want to ask, “Where have you seen or heard about us? Select all that apply.” That measures several remembered contacts. It does not tell you their order.
Keep those results separate. With multiple selections, channel percentages can add up to more than 100%. Don't describe them as each channel's share of all leads.
TrackLeadSource
See the story behind each lead.
Keep source, campaign, and journey context together when a visitor becomes a lead.
7-day free trial. No card required
Where to put the question
For an enquiry form, an optional question near the end is a reasonable starting point. Ask for the details needed to respond first. Don't select a source by default: an untouched answer could look like a real choice.
You can also ask after submission or during a sales conversation. That keeps the initial form shorter, but some people may never answer. Make sure any later answer reaches the same lead record.
There is no single placement that works best for every business. Check how people use your form and whether the extra question gets useful answers.
Before using the data, submit a test enquiry. Try a normal choice, Other with a written answer, and no answer. Check the saved form entry and your CRM or spreadsheet. The selected choice and any extra text should arrive together. A skipped question should stay blank.
Keep the question wording consistent when comparing periods. If you change from “first hear about us” to “find us today,” note the date. A change in results may reflect the new question.
People forget. Keep the recorded source too
A person may remember your name but forget where they first saw it. They may choose Google because that was the last step before contacting you. Someone else may remember a friend's recommendation but have no idea which ad they clicked later.
Automatic source tracking can preserve available visit details as someone browses. It doesn't need the person to remember or type them into a survey. When those details are passed with the form, you can keep them with the enquiry.
That is the problem TrackLeadSource is built around: keeping source, campaign, and journey context with form leads, so you can compare channels with customer outcomes.
But tracking cannot see everything. A conversation with a friend leaves no website visit to record. A visit on another device may not connect to the same journey. Some visits provide no identifiable source. The survey can add context that the recorded data lacks.
| Method | What it tells you | What to watch for |
|---|---|---|
| Survey answer | Where the person remembers hearing about you | Memory, unclear choices, and skipped answers |
| Recorded source | Available evidence from website visits, such as a tagged campaign link | Missing visit data and journeys that cannot be connected |
| Sales outcome | Whether the enquiry became a customer | Someone still needs to keep the outcome up to date |
For the practical capture method, see how to track lead source in website forms. The important part is saving the data with the individual lead, where you can use it later.
Two different answers can both be useful
Imagine this journey: a colleague recommends a business. The person later clicks a Google ad and submits a quote request.
An illustrative CRM record could look like this:
| Information | Saved value |
|---|---|
| Survey answer | Recommendation from someone |
| Survey detail | A colleague mentioned you |
| Recorded source | |
| Recorded medium | Paid search |
| Campaign | Office move |
| Sales outcome | Became a customer |
These are example labels and values, not a product screenshot or required field names.
Don't overwrite the recommendation with Google. Don't replace the recorded Google visit with the recommendation either. One tells you about remembered discovery. The other tells you about a recorded visit before the enquiry.
Store them separately and label reports clearly. If you need to decide how different recorded visits receive credit, read the guide to lead attribution models.
Read the results without hiding the missing answers
Start by checking how many people answered. A large share of one option among respondents may represent a much smaller share of all enquiries.
Here is an illustrative month:
| Enquiries | Count |
|---|---|
| Total enquiries | 100 |
| Answered the source question | 60 |
| Left it blank | 40 |
| Selected recommendation, within the 60 answers | 15 |
Recommendations account for 25% of respondents: 15 divided by 60. They account for 15% of all enquiries: 15 divided by 100. You do not know the remembered source of the 40 people who skipped the question.
Keep “I don't remember” separate from a blank response, too. One is an answer; the other is missing. Neither proves the visitor came directly to your website.
Then look at what happened to those leads. Which were suitable? Which became customers? Compare enquiries from the same period and allow time for sales to close. A channel with many new enquiries may still have deals in progress.
Survey counts alone are not a reason to increase a budget. Check customer outcomes and costs before making that decision. A channel can bring plenty of enquiries and few customers.
Review the written answers as well. If people repeatedly name the same directory under Other, consider giving it a choice of its own. Keep the original text when grouping similar answers, so you can check the meaning later.
Common questions
Only if your audience can make that distinction reliably. Many visitors will simply remember Google. You can keep that survey answer broad and use available campaign data to distinguish paid visits. Do not turn a person's “Google” answer into “organic search” without evidence.
Open text lets people describe something you did not expect. A dropdown makes answers easier to group. If you know little about how people find you, start with optional text and review the responses. A short list with an Other field is useful once common answers become clear.
It can be, especially if people already mention it. “ChatGPT or another AI tool” with an optional detail field keeps the wording simple. A survey answer does not prove someone clicked an AI referral link. They may have read a recommendation and then searched for your business elsewhere.
If you already saved their discovery answer, avoid asking again on every enquiry. If your form cannot distinguish returning customers, let them skip it. “I'm already a customer” can explain a new enquiry, but it does not tell you how they originally found you.
Start with a question you can use
Choose the moment you want to ask about. Give people clear choices and a way to say they don't remember. Check that the answer reaches the saved enquiry.
Keep recorded source data alongside it. When you later see which enquiries became customers, you will have more context for deciding where your marketing effort belongs.