Table of contents
Key takeaways
- A real recording shows the person behind the business. Whether that changes trust or conversion needs to be tested rather than inferred from unverified customer figures.
- Visitors look for proof that a business exists, responds, and has satisfied other humans before them.
- Start with a recording that answers a useful question on the page.
What does the evidence show about faces and text?
A face and voice can introduce a service provider, but the available customer examples do not substantiate an average enquiry increase.

Define the action you want to evaluate.
Track completed enquiries, bookings and purchases separately from session duration. The available examples do not establish a consistent increase in any of those measures.
A face can help identify the person behind a business. To build trust on your website, provide accurate contact and service information too. The available examples do not establish that recognition causes more enquiries.
Different audiences may respond differently. There is no verified average uplift across industries in the available evidence.
What makes visitors trust a website they have never seen before?
Visitors look for proof that a business exists, responds, and has satisfied other humans before them.

Make contact details easy to find and show that the information is current. Stanford’s 2002 credibility guidelines recommend clear contact information, a visible organization and content that has been updated or reviewed. They do not establish a universal three-second scan.
HTTPS protects the connection, but does not establish that the business is trustworthy. Use verifiable testimonials with permission and accurate context. The checked evidence does not establish a universal conversion ranking for testimonials, star ratings and security badges.
Video can introduce the team.
A team recording gives visitors a face and voice to associate with the business. Use clear audio and an accurate explanation. The available examples do not establish a neurological mechanism, an ideal production style or a predictable conversion benefit.
To build trust on website pages effectively, place the video where first time visitors land, not buried three clicks deep. The thumbnail should show the person looking toward the camera, not staring at a screen or reading from a script. A clear thumbnail can show who is speaking. It does not establish trust or a conversion advantage by itself.
Try a short recording in a quiet, well-lit setting. This is a practical starting point, not a documented conversion experiment.
Avoid autoplay with sound. Let visitors choose when to listen.
How to set up your first video greeting
Start with a recording that answers a useful question on the page. Allow time for recording, site access, configuration and testing.
What to say
You don’t need a teleprompter or a written monologue. Record a 15 second greeting that says who you are, what you help with, and what the visitor should do next. Keep it simple. Something like: “Hi, I’m Sarah from the support team. If you have questions about sizing, tap the button below and I’ll send you our fit guide.” Make the offer specific and keep the recording clear.
Where to place it
Placement determines whether anyone sees it.
If you want to build trust on your website, place the greeting where it answers a relevant question. Check spacing on mobile and desktop so it does not cover the main action. The available evidence does not establish a universally best placement.
Check the display controls for the format you use. CompleteGreet documents timed, scroll and exit-intent triggers for supported Shoppable experiences, not every standard Bubble. Test the experience on desktop and mobile.
How to measure results
Track completed enquiries per eligible visitor in randomly assigned groups with and without the greeting where possible. A before-and-after baseline can show a change, but traffic, seasonality and other edits can explain it. Record session duration separately.
Don’t obsess over play rates. A visitor who watches 10 seconds and then fills out your form converted. A visitor who watched the full 60 seconds and bounced did not.
Check your thumbnail before launch. Choose a clear frame that identifies the speaker and fits the page. Evaluate it alongside the message and placement rather than assuming it matters more than video quality.
Upload your file, set the position and follow the installation guide. CompleteGreet documents a header snippet or the Shopify app. Test the player and intended action before publishing.
Bentin and colleagues’ face-perception study reported an N170 response around 172 milliseconds. It concerned face detection, not the time needed to trust a business. It did not compare website conversions from video and text.
Use the research as background, then test the message and outcome on your own website.
Willis and Todorov’s 2006 experiments found that people formed trait impressions after brief exposure to unfamiliar faces. Those judgments were not a test of their accuracy, a positive trust response or a website conversion benefit. Rapid impressions do not establish that faces outperform written information.
Test the greeting against your current experience using the same audience and goal. The available evidence does not substantiate an engagement or enquiry advantage over a static header.
For a service page, count completed bookings as well as video interactions. The customer examples do not establish a booking increase or prove a transfer of trust from face to brand.
Video files and player scripts use network and processing resources. Compression and format selection can help, but no universal 2 MB threshold or sub-second requirement establishes a trust benefit. Measure the loading behavior of your actual page.
Test whether the greeting answers a question on the page and supports the intended action. Widget interactions alone do not establish a conversion lift against a static header.
The available evidence does not establish a 72-hour result or an automatic dashboard measure of causal uplift. Set up a baseline or comparable control and assess completed actions with visitor counts.
Common questions
Do faces on websites convert better than text?
A face and voice provide information about a person that plain text does not. They do not guarantee trust or conversion. Test whether the recording answers visitors’ questions and helps them complete the intended action.
How much can video greetings increase inquiries and conversions?
There is no verified typical uplift in the customer examples reviewed here. Define the goal, count eligible visitors and completed actions, and compare equivalent groups before attributing a change to video.
What should you say in a 30 second video greeting?
State your name, role and the problem you help with, then offer a next step. A fictional example is: ‘I am Sarah. I help contractors with their books. If that sounds useful, use the calendar below.’ This is sample wording, not a documented customer result.
Where should you place a video greeting on your website?
Consider pages where visitors need an explanation before acting, such as pricing, services or complex products. Test the placement rather than assuming it has the strongest return.
Do video widgets slow down your website?
Loading behavior depends on the widget, video files and page configuration. Test your own setup with Google PageSpeed Insights before and after adding the video, and assess conversion separately. A performance benefit or an offsetting conversion gain is not guaranteed.
Which businesses benefit most from video greetings?
Video can introduce a service provider or explain a complex product. Its value depends on the visitor’s question and the business goal; the available examples do not establish which industries get the strongest returns.
How do you measure if your video greeting is actually working?
Track completed enquiries or purchases per eligible visitor on the affected page. Use a control group where possible, or account for traffic and seasonal differences in a before-and-after comparison. Thirty days alone does not establish statistical confidence.
