# Self-Reported Attribution for SaaS

> How SaaS companies can use self-reported attribution to understand how buyers discover them, including question design, placement, categorization, analysis and combining it with software attribution.

Source: https://comppound.com/blog/self-reported-attribution-for-saas (published 2026-09-02)

Self-reported attribution means asking customers how they found you. Compared with tracking pixels, UTM parameters and attribution software, it can seem almost too simple to be useful. But it captures something software consistently misses: the conversations, podcasts, communities and recommendations that actually made someone aware of your product before they ever clicked a trackable link.

For SaaS companies, especially those investing in content, community and brand, self-reported attribution is one of the most useful sources of marketing insight available, and one of the cheapest. This guide explains why software attribution misses so much, how to ask the question, where to ask it, how to categorize and analyze the answers, and how to combine them with your other data without treating any single view as the whole truth.

## Why software attribution misses so much

Software attribution relies on trackable clicks. It can tell you that someone arrived from a search result, a tagged email or a paid ad. It struggles with almost everything that happens before that click.

Word of mouth is the biggest gap. When a colleague recommends your product in a meeting or a peer mentions it over lunch, there is no click to track. Private communities, such as Slack groups and invite-only forums, create a similar gap, because links shared there often appear as direct traffic or do not get clicked at all. Podcasts and videos usually lead listeners to search for your brand later, so the eventual visit is credited to search. Social posts that people see but do not click shape awareness without leaving a trace. AI assistants increasingly mention products in answers, often without a link the user follows. And privacy tools, browser restrictions and consent choices block tracking for a growing share of visitors.

Consider a typical path. A buyer hears about you on a podcast, sees a founder's LinkedIn post a week later, asks a colleague whether they have used you, and then searches your brand name on Google and signs up. Your analytics credits Google. The buyer, if asked, credits the podcast. Both are partially right, but only one of them tells you where demand was created.

## The basic question

Ask one open question at a key point in the conversion process:

> How did you hear about us?

Small variations can change the answers you get. "Where did you first hear about [Product]?" pushes people toward the earliest touchpoint they remember. "What made you decide to check us out?" captures the trigger that prompted action, which may be different. Choose the version that matches what you most want to learn, and keep it consistent over time so the data remains comparable.

## Free text or dropdown

The format of the question matters. A free text field captures detail, such as the name of a specific podcast, newsletter or colleague. That detail is often the most actionable part of the answer. The trade-off is that free text requires categorization before you can analyze it, and some answers will be vague or unhelpful.

A dropdown is easy to analyze and gives tidy charts, but it loses detail and can bias answers. People tend to pick from the options listed, so a channel you forgot to include will be underreported, and "Other" becomes a catch-all that tells you little. Dropdowns also encourage quick, careless selections.

A common compromise is a free text field, or a dropdown followed by an optional "Tell us more" field. That gives you structured data for reporting and specific details for decisions. Whichever you choose, include options or prompts for the channels that software tends to miss, such as podcasts, communities, AI assistants and recommendations from colleagues.

## Where to ask

Asking on the signup form captures nearly everyone, which gives you the most complete data. The trade-off is added friction at the moment you least want it, so keep the field optional or very short if you ask there.

An onboarding survey shown after account creation has lower friction, because the user has already committed. Response rates can still be good if the survey is short and appears before the user gets deep into the product.

A demo request form is a strong place to ask in sales-led motions, because the answer gives sales useful context before the first call. Knowing that a prospect heard about you from a specific community or podcast can shape the conversation.

Sales discovery calls offer a chance to ask conversationally and probe further. A salesperson can follow up on a vague answer and often learn the full story, such as the colleague who recommended you and the article they read first. Log these answers in the CRM so they become data rather than anecdotes.

Many teams combine these, asking in the onboarding or demo form and reinforcing it during sales calls for larger deals.

## Categorizing responses

For free text answers, create a consistent set of categories. Typical ones include search engine, AI assistant, LinkedIn, other social media, podcast, newsletter, word of mouth, community, review site, event, ads and blog or article. Adjust these to your business, but keep them stable once you start, because changing categories often makes trends impossible to read.

Within each category, also record the specific source when it is mentioned, such as the name of a podcast, newsletter or community. Those specifics are often the most actionable insight in the entire dataset. Knowing that "podcasts" matter is interesting. Knowing that one particular podcast keeps coming up tells you exactly where to invest next.

AI tools can help categorize large volumes of responses quickly. Spot-check the results regularly, because automated categorization can misread vague or ambiguous answers, and small errors in categorization add up over months.

## Getting useful answers

A few practical details improve the quality of responses considerably. Place the question where people are not rushing, which is why a short onboarding screen often works better than the last field of a long form. Explain briefly why you are asking, such as "This helps us understand where people find us", because a reason makes people more likely to give a real answer rather than typing "Google" to move on. Avoid making the field required on a signup form, since forced answers tend to be careless.

Make sure the answer flows somewhere people will see it. Store it as a field on the contact or account record in your CRM, and show it to whoever handles the first sales conversation. When sales and marketing both see the answers regularly, patterns get noticed sooner, and salespeople can confirm or correct vague answers during calls, which improves the data for everyone.

## Analyzing the data

Review the data monthly. Look at the share of responses by category, how those shares change over time, and whether the pattern differs between all signups and paying customers. That last comparison matters a lot. A channel that brings many signups but few paying customers may be less valuable than one that brings fewer but better-fit buyers. Also look for specific sources that keep appearing, such as a newsletter or community that shows up repeatedly.

The most useful comparison is often against software attribution. If self-reported answers say podcasts and word of mouth drive a large share of customers while analytics shows mostly direct and branded search traffic, that suggests your brand-building is working in ways analytics cannot see. In that situation the two views describe different moments in the same journey, so both can be accurate at once.

## A worked example (illustrative)

Consider a hypothetical SaaS company whose analytics showed most signups coming from direct traffic and organic search. The team assumed SEO was its main growth driver and was considering cutting time spent on other channels.

When they added a free text "How did you hear about us?" field to onboarding, a large share of customers mentioned a specific industry newsletter and the founder's LinkedIn posts. Many of those customers had later searched for the brand name, which is why analytics credited search. The team increased its investment in LinkedIn and sponsored the newsletter again, while continuing its SEO work. Without the self-reported field, they would have credited search for demand created elsewhere, and might have cut the channels that were actually creating it.

## Limitations

Self-reported attribution has real limitations, and it is important to be honest about them. People forget, simplify and misremember. Answers tend to reflect the most memorable touchpoint rather than the first one, so a vivid podcast episode may get credit over a blog post someone read months earlier. Response rates vary, especially if the field is optional, and the people who answer may not represent everyone. Some answers will be unhelpful, such as "Google" from someone who actually heard about you elsewhere and searched your name.

Treat self-reported attribution as one important view of how buyers find you, not the only truth. Its value lies in revealing what software cannot see, and it works best next to your software data rather than in place of it.

## Combine with other data

Use self-reported data alongside your other sources. First-touch and last-touch analytics show the trackable parts of the journey. UTM-tagged campaigns tell you which specific links people clicked. CRM notes from sales conversations add qualitative detail. And revenue and retention by source tell you which channels produce customers who stay and grow.

When several sources agree, you can act with confidence. When they disagree, that disagreement is useful. It usually points to a channel that is creating demand without being credited, or to a tracking gap that needs fixing. Investigate rather than simply choosing the number you prefer.

## Common mistakes

The most common mistake is asking the question but never analyzing the responses. Many companies have years of "How did you hear about us?" data sitting unused in a database. Another is using a dropdown that omits important channels, which guarantees they will be underreported. Teams also often fail to store responses with the customer record, which makes it impossible to compare sources by revenue or retention later. And changing categories so frequently that trends become meaningless undoes much of the value of collecting the data in the first place.

## The takeaway

Self-reported attribution is cheap, simple and surprisingly revealing. Ask one good question in a consistent place, categorize the answers carefully, record specific sources, and compare the results with your software attribution. Used this way, it often gives you the clearest view of the channels that shape buying decisions but never show up in analytics.

## Frequently asked questions

### What is self-reported attribution?

Asking customers directly how they heard about your company, usually through a form field or survey question.

### Is self-reported attribution reliable?

It is imperfect but valuable, especially for channels like word of mouth, podcasts and communities that analytics misses.

### Should the question be free text or a dropdown?

Free text captures more detail, while dropdowns are easier to analyze. Many teams combine both.

## About Comppound

Comppound runs SEO and AI search for B2B SaaS companies: a person sets the plan and reviews samples and every high-risk page, and AI speeds up production.

- [Get the free report](https://comppound.com/ai-visibility-report): When buyers ask ChatGPT about your category, are you in the answer?
- [Book a call](https://comppound.com/contact): Most of this works better when buyers can find you first