Understanding what drives a user to click.
Every time a user lands on a page they make rapid, often unconscious decisions about what to interact with. This empirical study investigates click behaviour across three call-to-action types (heading, button, image) at two positions (top, bottom) inside a blog-card interface, testing whether type or placement actually changes what people click.
Do the type or position of a call-to-action actually change what people click?
A controlled study, 11 participants, 432 clicks, 3 CTA types × 2 positions, analysed in SPSS.
No statistically significant difference, but habit wins: 66.7% preferred images, 50% defaulted to buttons.
Modern interfaces mix many clickable styles, so what do people actually click?
Users judge what is clickable from prior experience and visual cues, shapes, colours, affordances. Classic conventions (blue links, rectangular buttons) gave strong signals, but contemporary interfaces present many different clickable elements, so designers need updated, evidence-based guidance.
Which call-to-action, text heading, button, or image, receives the most clicks?
Does placing the CTA at the top vs. the bottom of a blog card change click behaviour?
…establish which CTA options users prefer, and whether their placement significantly affects click behaviour?
What prior research says about click behaviour.
Users bring expectations from past web use; recognisable labels and layouts that match mental models drive clicks.
Decorative stock imagery is ignored, but real photos are studied intensely; topic interest correlates with click likelihood.
CTA design and format directly alter ad recognition and downstream conversion behaviour.
Across 1,000+ actions by Taobao users, click choices follow predictable, retrieval-based patterns.
Two hypotheses, formally stated.
All CTA elements are equally effective in prompting clicks (p > 0.05).
Visual elements, buttons and images, are more effective than text headings.
CTA position (top vs. bottom) has no significant impact on clicking behaviour.
Position has a minimal but measurable impact on clicking behaviour.
A controlled experiment with balanced counterbalancing.
Participants worked through a Figma prototype of blog cards, each carrying a heading, button and image CTA at either the top or bottom. A balanced Latin-square design across two groups removed order and learning bias.
Data collected
| Parameter | Description | Count |
|---|---|---|
| Participants tested | University students & staff · 11 valid after exclusion | 12 |
| Tested conditions | 3 CTA types × 2 positions | 6 |
| CTA types | Text heading · button · image | 3 |
| Clicks / participant | 6 for position + 6 for intermediate pages | 12 |
| Total recorded clicks | 12 × 2 conditions × 3 objects × 6 repetitions | 432 |
What 432 clicks revealed.
A Shapiro-Wilk test showed the data was non-normal (Button p<.001, Image p=.034, Heading p=.514), so I used non-parametric Mann-Whitney U tests throughout.
RQ1 · CTA type, mean clicks per element
| Comparison | U | Z | p-value | Decision |
|---|---|---|---|---|
| Button × Image | 529.5 | −0.033 | 0.974 | Accept H₀ |
| Button × Heading | 87.5 | −0.921 | 0.375 | Accept H₀ |
| Image × Heading | 62.5 | −0.989 | 0.322 | Accept H₀ |
RQ2 · Position, mean clicks by placement
But preference told a different story
“Used the buttons throughout. I knew I could have clicked the header or image but chose to continue to do what I learnt.”
What it means for designers.
Participants defaulted to rectangular buttons even when told images and text were clickable, learned signifiers win.
Even without significance at N=11, preference overwhelmingly favoured visual CTAs. Prioritise visual clarity over text-only links.
Top vs. bottom yielded near-identical click rates, contradicting prior work that placement drives behaviour. Once users learn where a CTA lives, they click it regardless of position.
What worked, and what I’d change.
- Rigorous SPSS non-parametric pipeline (Shapiro-Wilk + Mann-Whitney U)
- Balanced Latin-square counterbalancing eliminated order bias
- Ishihara colour-blindness pre-screening protected data integrity
- Test CTAs across different interface designs and contexts, not only blog cards
- Increase the sample beyond N = 11 for greater statistical power
- Recruit beyond Figma-literate university students and staff for broader generalisability
Sources & further reading.
Read the full research paper ↗- Burnett, J. C. (2014). Web Design Factors That Influence User Behavior.
- Chen, T.-Y., Yeh, T.-L., & Chang, C.-I. (2018). How different advertising formats and calls to action on videos affect advertising recognition and consequent behaviours. The Service Industries Journal, 40(5-6), 358-379.
- Li, D., Tang, Z., & Zhao, N. (2023). How does users’ interest influence their click behavior? Frontiers in Psychology, 14.
- Lohtia, R., Donthu, N., & Hershberger, E. K. (2003). The Impact of Content and Design Elements on Banner Advertising Click-Through Rates. Journal of Advertising Research, 43(4), 410-418.
- Lu, H., Zhang, M., & Ma, S. (2018). Between Clicks and Satisfaction: Multi-Phase User Preferences and Satisfaction for Online News Reading. ACM SIGIR.
- Nielsen, J. (2010). Photos as Web Content. Nielsen Norman Group.
- Nielsen Norman Group. (n.d.). Beyond Blue Links: Making Clickable Elements Recognizable.
- Qin, J., et al. (2020). User Behavior Retrieval for Click-Through Rate Prediction. SIGIR 2020, 1440-1449.
- Shapiro, S. S., & Wilk, M. B. (1965). An analysis of variance test for normality (complete samples). Biometrika, 52(3/4), 591-611.