The Two-Tool Setup That Actually Improves Your Portfolio Three weeks after launching your portfolio you have 240 visits. You have no idea which of your eight projects pulled them in. You have no idea...
The Two-Tool Setup That Actually Improves Your Portfolio
Three weeks after launching your portfolio you have 240 visits. You have no idea which of your eight projects pulled them in. You have no idea where they came from. You have no idea how many reached your contact page. The portfolio is performing or it is not, but you cannot tell which.
This is the analytics problem most indie creators ignore until year two, when they realize they have been adding new projects without ever cutting the weak ones. By then the catalog is bloated and the data needed to clean it up was never collected.
This guide expands the analytics section referenced in the Indie Creator Portfolio Playbook. The pillar mentions the quarterly review ritual. This one covers the exact tools and metrics that make the review worth doing.
A blog needs to track read-through rate and email signups. An ecommerce site needs to track cart abandonment and refund rate. A portfolio needs something different.
The questions a portfolio analytics setup should answer:
Which projects pull viewers in? Where do those viewers come from? Do they reach your contact page or your services list? How long do they stay on individual projects? Are they viewing on mobile or desktop?
You do not need a 47-dashboard analytics tool to answer these. The right setup is the simplest one that surfaces these five answers reliably.
If you host on Devdazzle, the Showcase profile gives you per-piece view counts surfaced on each portfolio item. That is your baseline. You can see which of your pieces accumulated the most views over time, which tells you which work is pulling attention.
Similar platforms expose comparable per-piece view dashboards. Industry-standard portfolio platforms typically include view counts and impression data on paid tiers. Playable-build hosts show per-page view counts for game projects. Interactive 3D viewer platforms surface view plus embed stats per model.
Platform-built analytics are useful for "which piece is popular" but limited in two specific ways. They typically do not show referrer (where the viewer came from). They do not show time-on-page or contact click-through rate.
For deeper data, you either rely on the platform expanding what they show you, or you add a self-hosted analytics layer on top.
If you have your own domain (an anchor site or a full self-hosted portfolio), three analytics tools cover most needs.
Plausible is privacy-respecting, paid, with no cookie banner required in most jurisdictions. The dashboard is simple and the pricing runs $9 to $19 a month depending on traffic volume. Good for creators who want clean data without complexity.
Umami is open-source, self-hosted if you host it yourself (free), or paid cloud. The feature set is similar to Plausible. Best for creators who already have a server.
Google Analytics 4 is free, comprehensive, and complex. It requires a cookie consent banner in most jurisdictions. It shows everything but takes weeks to learn the interface.
The choice rarely matters for the first year. Pick one and configure it to track these four things: page views per project, time on page per project, click-throughs to your contact page, and referral source for each visitor. Skip the rest of the defaults.
After three months of data, four metrics tell you what to fix.
Time-on-Page per Project
For context, the cross-industry average session duration is around 2 minutes and 17 seconds in 2025 (2025 web benchmarks), with long-form content pages generally pulling longer durations than short-form. For portfolio project pages, the goal is enough time for a viewer to actually look at the work and read the case study, not a specific number.
After three months of data, you will see clearly which pieces hold attention and which do not. The pieces with the highest time-on-page are not always the prettiest. Sometimes a piece with a strong written case study outperforms a more polished visual that lacks context. That data is gold for deciding what to keep.
Contact Page Conversion Rate
If 1,000 people visit your portfolio and none reach the contact page, your pieces are not strong enough, or your contact link is hidden, or there is no clear call-to-action.
Industry-typical cold-traffic conversion rates for B2B service sites land in the low single digits as a benchmark. Anything well above that range for a portfolio site usually indicates either highly qualified inbound traffic or a strong portfolio-to-CTA flow. Track your own number quarterly rather than chasing a specific benchmark.
Referral Source Distribution
Track where your visitors come from. If 80 percent come from one channel, you are over-indexed and exposed to that channel's algorithm risk. If your Discord outreach should be producing visits but you see zero referrals from Discord URLs, the outreach is not landing.
Use UTM tags on outreach links. A URL like
Mobile vs Desktop Ratio
In 2025, mobile bounce rates run higher than desktop on average across industries (mobile around 58 to 60 percent, desktop around 48 to 50 percent in cross-industry data). If a substantial share of your traffic is mobile but your portfolio's mobile layout breaks below 768px, you have a measurable problem. Most analytics tools show this split by default.
A common mistake: building the portfolio on a 27-inch monitor, optimizing for desktop, and never checking that mobile bounce rate is acceptable. Watch mobile bounce as a separate metric from desktop.
A few metrics most analytics dashboards prominently show but that mislead portfolio creators.
Total Visits
A vanity metric. 10,000 visits with 0 contact submissions is worse than 200 visits with 5 contact submissions. Track conversion, not raw volume.
Session Duration
Misleading for portfolios. A viewer who opens your portfolio, looks at three projects, and closes the tab might have decided in 90 seconds that they want to hire you. Session duration looks low even though the engagement was high.
Bounce Rate
The default definition (visitor leaves after one page) does not work for portfolio sites. Many portfolio visits are single-page by nature. Look at one project, close. That is normal and does not signal a problem.
Browser and OS Breakdown
Unless you are building experimental web tech, you do not need to know your visitor split between Chrome and Edge. Skip the report.
Every three months, sit down with the data for one hour.
Sort all project pages by time-on-page. The bottom 20 percent of projects either get cut or get rewritten with better case study text. The top 20 percent get promoted on social media and pitched as featured pieces in cold outreach.
Look at referral sources. Which channels produced visits this quarter? Which channels you spent time on produced no traffic? Adjust next quarter's outreach budget accordingly.
Check the mobile vs desktop split. If mobile has drifted up to 80 percent, verify your mobile layout still works on real devices.
Most creators never do this review. Their portfolios accumulate weak pieces for years without revision. The creators who run the cycle every quarter compound the catalog quality over time. After two years of quarterly reviews, the portfolio is unrecognizable from where it started.
Analytics in month one tell you almost nothing. The sample size is too small. Three months is the minimum useful data window.
The compounding starts when you have a year of quarterly reviews behind you. By then you know which channels produce real traffic, which pieces convert visits to inquiries, and which mobile breakages cost you the most. The portfolio is no longer guessing.
Analytics is the difference between iterating with data and iterating with hope. Pick one tool, configure four metrics, run the quarterly review. That is the entire setup.
Three weeks after launching your portfolio you have 240 visits. You have no idea which of your eight projects pulled them in. You have no idea where they came from. You have no idea how many reached your contact page. The portfolio is performing or it is not, but you cannot tell which.
This is the analytics problem most indie creators ignore until year two, when they realize they have been adding new projects without ever cutting the weak ones. By then the catalog is bloated and the data needed to clean it up was never collected.
This guide expands the analytics section referenced in the Indie Creator Portfolio Playbook. The pillar mentions the quarterly review ritual. This one covers the exact tools and metrics that make the review worth doing.
What Analytics Means for a Portfolio
A blog needs to track read-through rate and email signups. An ecommerce site needs to track cart abandonment and refund rate. A portfolio needs something different.
The questions a portfolio analytics setup should answer:
Which projects pull viewers in? Where do those viewers come from? Do they reach your contact page or your services list? How long do they stay on individual projects? Are they viewing on mobile or desktop?
You do not need a 47-dashboard analytics tool to answer these. The right setup is the simplest one that surfaces these five answers reliably.
Platform-Built Analytics
If you host on Devdazzle, the Showcase profile gives you per-piece view counts surfaced on each portfolio item. That is your baseline. You can see which of your pieces accumulated the most views over time, which tells you which work is pulling attention.
Similar platforms expose comparable per-piece view dashboards. Industry-standard portfolio platforms typically include view counts and impression data on paid tiers. Playable-build hosts show per-page view counts for game projects. Interactive 3D viewer platforms surface view plus embed stats per model.
Platform-built analytics are useful for "which piece is popular" but limited in two specific ways. They typically do not show referrer (where the viewer came from). They do not show time-on-page or contact click-through rate.
For deeper data, you either rely on the platform expanding what they show you, or you add a self-hosted analytics layer on top.
Self-Hosted Analytics Tools
If you have your own domain (an anchor site or a full self-hosted portfolio), three analytics tools cover most needs.
Plausible is privacy-respecting, paid, with no cookie banner required in most jurisdictions. The dashboard is simple and the pricing runs $9 to $19 a month depending on traffic volume. Good for creators who want clean data without complexity.
Umami is open-source, self-hosted if you host it yourself (free), or paid cloud. The feature set is similar to Plausible. Best for creators who already have a server.
Google Analytics 4 is free, comprehensive, and complex. It requires a cookie consent banner in most jurisdictions. It shows everything but takes weeks to learn the interface.
The choice rarely matters for the first year. Pick one and configure it to track these four things: page views per project, time on page per project, click-throughs to your contact page, and referral source for each visitor. Skip the rest of the defaults.
The Four Metrics That Matter
After three months of data, four metrics tell you what to fix.
Time-on-Page per Project
For context, the cross-industry average session duration is around 2 minutes and 17 seconds in 2025 (2025 web benchmarks), with long-form content pages generally pulling longer durations than short-form. For portfolio project pages, the goal is enough time for a viewer to actually look at the work and read the case study, not a specific number.
After three months of data, you will see clearly which pieces hold attention and which do not. The pieces with the highest time-on-page are not always the prettiest. Sometimes a piece with a strong written case study outperforms a more polished visual that lacks context. That data is gold for deciding what to keep.
Contact Page Conversion Rate
If 1,000 people visit your portfolio and none reach the contact page, your pieces are not strong enough, or your contact link is hidden, or there is no clear call-to-action.
Industry-typical cold-traffic conversion rates for B2B service sites land in the low single digits as a benchmark. Anything well above that range for a portfolio site usually indicates either highly qualified inbound traffic or a strong portfolio-to-CTA flow. Track your own number quarterly rather than chasing a specific benchmark.
Referral Source Distribution
Track where your visitors come from. If 80 percent come from one channel, you are over-indexed and exposed to that channel's algorithm risk. If your Discord outreach should be producing visits but you see zero referrals from Discord URLs, the outreach is not landing.
Use UTM tags on outreach links. A URL like
yourportfolio.com?utm_source=linkedin&utm_campaign=cold_2026q3 lets the analytics tool group visits by campaign so you can compare LinkedIn DMs against Reddit posts against cold emails.Mobile vs Desktop Ratio
In 2025, mobile bounce rates run higher than desktop on average across industries (mobile around 58 to 60 percent, desktop around 48 to 50 percent in cross-industry data). If a substantial share of your traffic is mobile but your portfolio's mobile layout breaks below 768px, you have a measurable problem. Most analytics tools show this split by default.
A common mistake: building the portfolio on a 27-inch monitor, optimizing for desktop, and never checking that mobile bounce rate is acceptable. Watch mobile bounce as a separate metric from desktop.
Metrics to Ignore
A few metrics most analytics dashboards prominently show but that mislead portfolio creators.
Total Visits
A vanity metric. 10,000 visits with 0 contact submissions is worse than 200 visits with 5 contact submissions. Track conversion, not raw volume.
Session Duration
Misleading for portfolios. A viewer who opens your portfolio, looks at three projects, and closes the tab might have decided in 90 seconds that they want to hire you. Session duration looks low even though the engagement was high.
Bounce Rate
The default definition (visitor leaves after one page) does not work for portfolio sites. Many portfolio visits are single-page by nature. Look at one project, close. That is normal and does not signal a problem.
Browser and OS Breakdown
Unless you are building experimental web tech, you do not need to know your visitor split between Chrome and Edge. Skip the report.
The Quarterly Review Ritual
Every three months, sit down with the data for one hour.
Sort all project pages by time-on-page. The bottom 20 percent of projects either get cut or get rewritten with better case study text. The top 20 percent get promoted on social media and pitched as featured pieces in cold outreach.
Look at referral sources. Which channels produced visits this quarter? Which channels you spent time on produced no traffic? Adjust next quarter's outreach budget accordingly.
Check the mobile vs desktop split. If mobile has drifted up to 80 percent, verify your mobile layout still works on real devices.
Most creators never do this review. Their portfolios accumulate weak pieces for years without revision. The creators who run the cycle every quarter compound the catalog quality over time. After two years of quarterly reviews, the portfolio is unrecognizable from where it started.
The Compounding Curve
Analytics in month one tell you almost nothing. The sample size is too small. Three months is the minimum useful data window.
The compounding starts when you have a year of quarterly reviews behind you. By then you know which channels produce real traffic, which pieces convert visits to inquiries, and which mobile breakages cost you the most. The portfolio is no longer guessing.
Analytics is the difference between iterating with data and iterating with hope. Pick one tool, configure four metrics, run the quarterly review. That is the entire setup.