Alex is standing in the middle of a city with a modern phone, a full battery and a large 5G symbol at the top of the screen.

He opens a website.

It takes several seconds.

A video starts in low resolution.

Then a file upload almost stops completely.

Alex looks at the signal indicator again.

Five bars.

5G.

Everything appears perfect.

So why does the connection feel worse than the 4G connection he used yesterday?

Because the small symbols at the top of a smartphone hide most of what actually determines mobile internet performance.

A phone can display 5G and still have a disappointing connection.

It can show fewer signal bars and work extremely well.

And in some situations, switching from 5G to 4G can actually improve the experience.

The useful question is therefore not:

„Do I have 5G?“

It is:

„Which part of the mobile connection is currently limiting me?“

5G Is a Radio Technology, Not a Speed Guarantee

The 5G icon tells Alex something important.

His phone is connected to, or participating in, a network using 5G technology.

But it does not tell him:

• how much spectrum is available,
• how many other users are sharing the cell,
• how strong the usable radio signal really is,
• whether the network is using low-band, mid-band or high-band frequencies,
• whether the local base station has enough backhaul capacity,
• whether the server he is accessing is slow,
• or whether his operator is limiting the connection.

That is why two people can both see 5G and experience completely different speeds.

One might download at several hundred megabits per second.

Another might struggle to reach 20 Mbit/s.

The network label is only the beginning of the explanation.

Not All 5G Frequencies Behave the Same Way

One of the biggest differences comes from frequency.

Mobile operators use different parts of the radio spectrum to provide coverage and capacity.

For a simplified comparison, think of three broad groups.

Low-band

Lower frequencies travel relatively far and penetrate buildings better.

That makes them useful for broad geographic coverage.

But the amount of spectrum available may be limited.

A low-band 5G connection can therefore provide excellent coverage without producing spectacular speed.

Mid-band

Mid-band spectrum often provides one of the most useful compromises.

It can offer significantly more capacity than many low-band deployments while still covering practical distances.

This is where many of the impressive everyday 5G speed improvements come from.

High-band or millimeter wave

Very high frequencies can offer enormous capacity over relatively short distances.

But signals are easier to block.

Buildings, walls, trees and even the geometry of a street can matter.

A connection that is extremely fast near the antenna may degrade quickly as conditions change.

So when Alex sees the same 5G icon in two places, the underlying radio connection may be completely different.

The phone’s status bar does not explain that.

Five Signal Bars Do Not Mean Five Bars of Internet Capacity

Alex assumes his five bars mean the connection should be fast.

That is another common misunderstanding.

Signal bars are primarily a simplified representation of radio signal conditions.

They do not measure how much total internet capacity is available to Alex.

Imagine a restaurant with excellent Wi-Fi coverage.

The access point is directly above the table.

Signal strength is perfect.

But 200 people are all trying to stream video through the same internet connection.

The problem is not reaching the access point.

The problem is shared capacity beyond it.

A mobile network works differently in detail, but the principle is useful.

Alex may have a strong radio connection to the cell tower while competing with hundreds of other devices for available resources.

The bars can remain full while performance collapses.

A Cell Tower Is Shared Infrastructure

Mobile capacity is not dedicated to one person.

A base station serves many users.

Suppose, purely as an illustration, a sector has 1,000 Mbit/s of usable capacity available at a particular moment.

If only 20 users are actively consuming substantial data, the network may feel extremely fast.

If 300 users become active at the same time, things change.

That does not mean every user receives exactly:

1,000 ÷ 300 = 3.3 Mbit/s

Real scheduling is much more sophisticated than equal division.

Users have different signal quality, traffic requirements, network priorities and radio conditions.

But the basic point remains:

capacity is shared.

This explains why the same location can feel excellent at 06:00 and terrible at 18:00.

Nothing may have changed in Alex’s phone.

Nothing may have changed in the distance to the tower.

What changed is the number and behavior of other users.

The Stadium Problem

The effect becomes obvious at concerts, stadiums, festivals and busy transport hubs.

Alex attends a football match.

His phone displays strong 5G.

Before kickoff, a photo uploads almost instantly.

During halftime, tens of thousands of people take out their phones.

Messages, videos, photos and social media requests all increase at roughly the same time.

Alex’s signal still looks excellent.

But a photo now takes 40 seconds to send.

This is a capacity problem.

More antennas, more spectrum and better network design can reduce it, but the 5G icon alone cannot tell Alex whether enough capacity exists for the crowd around him.

The Tower Still Needs a Fast Connection to the Rest of the Internet

There is another part of the journey that smartphone users rarely see.

The phone does not communicate directly with Netflix, Google, a cloud service or the website Alex is opening.

A simplified path looks like this:

Phone → radio network → base station → operator network → internet → destination server

The connection from the mobile site into the operator’s network is often called backhaul.

It may use fiber, microwave links or other infrastructure.

Imagine the radio side could theoretically handle 800 Mbit/s for Alex.

But the local site’s connection toward the rest of the network is heavily loaded.

The fast radio link does not solve that bottleneck.

It is similar to connecting a laptop to a 2.5-gigabit home router while the home’s internet connection itself provides only 100 Mbit/s.

The local link is fast.

The end-to-end connection is still limited elsewhere.

Download Speed Is Only One Part of the Experience

Alex runs a speed test.

Result:

Download: 280 Mbit/s

That looks excellent.

Yet a video call still feels unstable.

The reason is that download throughput is not the only metric that matters.

Consider four different measurements:

Metric	What it influences
Download speed	Receiving large files, streaming, browsing
Upload speed	Sending files, cloud backup, video calls
Latency	Responsiveness and delay
Packet loss / stability	Reliability of real-time connections

A connection can have excellent download speed and still produce poor real-time performance.

Suppose Alex has:

280 Mbit/s download

but

6 Mbit/s upload

with rapidly changing latency.

Downloading a large app may be fast.

Uploading a 2 GB video could be frustrating.

A video meeting might become unstable if the connection experiences sudden delay or packet loss.

Calling a network „fast“ therefore hides several separate behaviors.

280 Mbit/s Does Not Make Every Website Open Instantly

Alex expects 280 Mbit/s to make everything instantaneous.

But many everyday tasks require surprisingly little raw bandwidth.

A website might transfer only a few megabytes.

The delay could come from:

• DNS resolution,
• server response time,
• advertising scripts,
• third-party trackers,
• app processing,
• network latency,
• or a badly optimized website.

Once a connection is already comfortably fast enough for the data involved, adding more bandwidth can produce little visible improvement.

Suppose a webpage transfers 3 MB.

At 30 Mbit/s, the theoretical raw transfer time is roughly:

3 MB × 8 = 24 megabits

24 ÷ 30 ≈ 0.8 seconds

At 300 Mbit/s:

24 ÷ 300 ≈ 0.08 seconds

The theoretical difference is about:

0.72 seconds

And even that ignores protocol overhead and all the other work required to build the page.

If the web server itself waits 1.5 seconds before responding, Alex’s tenfold increase in connection speed does not remove that delay.

That is one reason enormous speed-test numbers do not always make phones feel ten times faster.

Why 4G Can Sometimes Beat 5G

Alex performs an experiment.

With 5G enabled:

38 Mbit/s download

He switches temporarily to 4G.

Now:

74 Mbit/s download

That seems backwards.

But several explanations are possible.

The 5G layer may be congested.

The phone may be connected to a weak or less favorable 5G carrier.

The 4G network may have more usable capacity at that location.

Network configuration and carrier aggregation can also affect the result.

The labels 4G and 5G describe generations of mobile technology.

They do not establish a rule that every individual 5G connection must outperform every individual 4G connection.

In a well-designed modern network, 5G can create enormous improvements in capacity and efficiency.

But Alex is not using an abstract network.

He is using one specific cell, at one specific time, under one specific set of conditions.

That distinction matters.

Before Alex Blames 5G, He Tests the Building

The next day, Alex notices something strange.

His connection is poor at his desk.

He walks toward the window.

Suddenly the speed improves dramatically.

Nothing changed in his mobile plan.

The network did not receive an upgrade during those ten seconds.

What changed was the radio environment between his phone and the network.

Buildings can weaken mobile signals. The effect depends on frequency, wall construction, insulation, coated glass, metal structures and the position of the phone relative to the serving cell.

This creates situations where a phone appears to have reasonable coverage but the quality of the radio link is much worse than it is outdoors.

Alex tests three locations:

At his desk:

42 Mbit/s

Near the window:

185 Mbit/s

Outside:

310 Mbit/s

Those numbers do not prove that every building will behave this way.

They tell Alex something much more useful:

his phone and mobile plan are capable of much better performance than he receives at his desk.

The location inside the building has become an important suspect.

Moving a Few Meters Can Sometimes Matter More Than Buying a New Phone

This is particularly useful when a smartphone is being used as a hotspot.

Imagine Alex works from a laptop through his phone while staying in a holiday apartment.

At the sofa, the connection is unstable.

Instead of immediately blaming the operator, he places the phone near a window and leaves it there as a hotspot.

His laptop remains at the table.

The phone now has a better radio environment while Wi-Fi covers the short distance inside the apartment.

It costs nothing to test.

The same principle can apply to dedicated 4G or 5G routers. Placement can matter enormously, particularly when the available signal is marginal.

Before buying antennas, boosters, routers or another phone, Alex first asks the cheapest diagnostic question:

Does the connection improve significantly somewhere else in the same building?

If yes, he has learned something.

A Phone in Motion Has Another Problem

Alex experiences a different kind of instability on a train.

His phone is not communicating with one permanent tower for the entire journey.

As he moves, the network may need to transfer the connection between cells.

This process is generally called a handover or handoff.

Most of the time, users barely notice it.

But high speed changes the radio environment continuously.

A train can move through:

• strong coverage,
• weak coverage,
• tunnels,
• cuttings,
• rural gaps,
• overloaded station areas,
• and cells operating on different frequency combinations.

A speed test at one moment therefore says surprisingly little about what Alex will experience ten kilometers later.

This is also why a video stream can work perfectly for twenty minutes and suddenly buffer even though the phone continues to display a mobile-network symbol.

The connection is not one fixed pipe.

It is a journey through changing network conditions.

His Mobile Plan Can Be the Bottleneck Too

Alex now has another possibility to check:

the tariff itself.

Two customers can use similar phones on the same operator’s network and still receive different performance because their plans are not necessarily identical.

Depending on the operator and tariff, there may be:

• maximum speed limits,
• different network access,
• data allowances,
• restrictions after a threshold,
• or other traffic-management conditions.

Suppose the network could provide Alex with 300 Mbit/s at his location.

If his plan is limited to 50 Mbit/s, replacing a 5G phone with a newer 5G phone will not remove a 50 Mbit/s tariff limit.

This gives him another useful diagnostic clue.

If repeated tests under excellent conditions keep stopping around almost the same number, he checks the terms of his mobile plan.

A ceiling created by a contract cannot be repaired by moving closer to a tower.

Dual-SIM Can Make Diagnosis More Interesting

Alex uses two SIMs.

One is personal.

The other belongs to his employer.

That gives him a useful comparison.

If both SIMs use different networks, he can test whether the problem follows the phone or the operator.

Suppose at the same desk, using the same device:

SIM A: 24 Mbit/s

SIM B: 160 Mbit/s

That does not automatically prove that Network B is universally better.

At another address, the result could reverse.

But it strongly suggests that Alex’s original problem is not simply:

„My phone is too slow.“

The same hardware has just demonstrated much higher performance under nearly identical local conditions.

For someone choosing a mobile provider, this also explains why national coverage claims cannot answer every personal question.

The network that performs best where Alex actually lives, works and travels can matter more than the network with the most impressive headline statistic.

A VPN Adds Another Route to the Journey

Alex normally uses a VPN on his phone.

He switches it off temporarily and repeats a test.

The connection becomes faster.

Again, this does not mean VPNs are inherently slow or that he should stop using one whenever performance matters.

A VPN changes the path and processing involved in his traffic.

Performance can depend on:

• the VPN provider,
• server location,
• server load,
• protocol,
• encryption overhead,
• routing,
• and the quality of the underlying mobile connection.

A badly chosen or overloaded VPN endpoint can therefore become the bottleneck even when 5G itself is performing well.

Alex tests rather than guesses.

If:

5G + VPN = 35 Mbit/s

and

5G without VPN = 240 Mbit/s

he has found an important clue.

If both tests remain around 35 Mbit/s, he continues looking elsewhere.

One Slow App Does Not Prove the Network Is Slow

There is an even simpler trap.

Alex says:

„My internet is slow.“

But perhaps only one application is slow.

He tries:

• several websites,
• a large file download,
• another streaming service,
• and a speed test.

Everything works normally except one app.

At that point, the mobile network becomes a less convincing explanation.

The service itself may be experiencing problems.

Its server may be overloaded.

The app may have a bug.

Its content delivery infrastructure may be having trouble in Alex’s region.

The distinction matters because changing network settings cannot repair a remote service.

Before diagnosing an entire mobile connection from one failed app, Alex checks whether the problem exists across several unrelated services.

Alex Uses a Five-Minute Diagnosis Instead of Guessing

He now has a practical sequence for the next time 5G feels slow.

Test 1: Change location

He moves toward a window or goes outside.

If performance changes dramatically, the radio environment is probably important.

Test 2: Test at another time

If the connection is excellent in the morning but consistently poor during busy evening hours, local congestion becomes a plausible explanation.

Test 3: Compare 5G and 4G

Alex temporarily tests both network modes where his phone and operator allow it.

He does not assume the newer label must win every test.

Test 4: Test more than one service

If only one website or app is slow, he investigates that service before blaming the network.

Test 5: Remove optional layers

For diagnosis, he can temporarily test without a VPN or other network-altering software.

The purpose is not to permanently disable useful privacy or security tools.

It is to isolate the variable.

Test 6: Check the tariff

He verifies whether his plan contains a maximum speed or other relevant restrictions.

Test 7: Compare another device or network

If available, another phone or SIM can be extremely informative.

The goal is not to produce a laboratory benchmark.

It is to discover which variable makes the problem disappear.

The Pattern Matters More Than One Speed Test

Suppose Alex records these results:

Test	Download
Desk, 18:30	28 Mbit/s
Window, 18:35	61 Mbit/s
Outside, 18:40	74 Mbit/s
Outside, next morning	285 Mbit/s

That pattern tells a story.

The improvement from desk to outside suggests the building matters.

The much larger improvement the following morning suggests congestion may matter too.

There may be more than one bottleneck.

That is common.

Real network problems do not have to fit neatly into one category.

Now consider another result:

Test	Download
Desk	48 Mbit/s
Window	49 Mbit/s
Outside	50 Mbit/s
Different time	49 Mbit/s

If Alex’s tariff is limited to 50 Mbit/s, the pattern suddenly makes sense.

Buying a new €1,000 smartphone to solve that problem would be a very expensive diagnostic mistake.

When a New Phone Actually Can Help

This does not mean hardware never matters.

Phones differ in modem generations, supported frequency bands, carrier aggregation capabilities, antenna design and other radio features.

An old device may lack support for frequencies or network capabilities that an operator now uses extensively.

A damaged device can also behave abnormally.

A modern phone may therefore provide meaningful improvements in some circumstances.

But Alex wants evidence before replacing it.

He asks:

Does another compatible phone on the same network perform substantially better at the same place and time?

If yes, hardware becomes a stronger suspect.

If both devices struggle in exactly the same location during exactly the same busy period, buying another handset may change very little.

The 5G Icon Is a Clue, Not a Diagnosis

Alex eventually stops using the symbol at the top of the screen as a performance score.

It was never designed to tell him the whole story.

His mobile internet experience is the result of a chain:

Phone → radio conditions → spectrum → cell capacity → operator network → backhaul → internet routing → destination service

And sometimes:

→ VPN or other additional processing

A bottleneck anywhere along that chain can dominate the experience.

That explains the apparent contradictions.

Five bars can feel slow.

Three bars can feel fast.

5G can lose to 4G in one particular place.

A 300 Mbit/s speed test can coexist with an annoying website.

A €1,000 phone can perform badly inside one building while a cheaper phone on another network works perfectly.

None of those observations breaks the rules of mobile networking.

They simply reveal how little information fits inside the tiny 5G icon.

The next time Alex’s phone says 5G while the internet crawls, he does not immediately restart the phone, blame the operator or start shopping for new hardware.

He changes one variable.

Then another.

Within a few minutes, he can often answer a far more useful question than:

„Why is my 5G slow?“

He can begin to answer:

„Where, exactly, is the bottleneck?“