Twenty Years of Bluetooth Fell Short of Its Promise — Failed Communication Technologies Tend to Fail the Same Way | QDaily Business History

The communication technology on every phone that almost nobody actually uses

Apple once made a Bluetooth headset. It shipped as an iPhone accessory, came with wireless charging, looked strange, cost a lot, and sold terribly.

It was not called AirPods, of course. It had the blunt name "Apple iPhone Bluetooth Headset" and arrived in 2007 alongside the original iPhone.

Unlike the original iPhone, which upended the global telecom market, this headset left no mark. Its greatest value today may be as a rare Apple collectible trading on eBay.

For more than a decade after the first Bluetooth headset arrived in 2000, single-ear Bluetooth pieces stayed unpopular. Wearing one — sometimes with a blinking light — made you look like a chauffeur or a secret-agent wannabe. Wired, in 2009, joked: "Maybe you watched too much RoboCop as a kid and figured decorating yourself with electronics was cool. Let us be clear: walking around with a Bluetooth earpiece in one ear is stupid. No exceptions."

There were other problems too: fiddly pairing, dropped connections, constant recharging.

Not even Apple under Steve Jobs could make people embrace Bluetooth headsets. That first attempt failed so badly that Apple waited nearly a decade to try again — with AirPods, which sold 35 million units last year.

AirPods now outsell every other Bluetooth headset combined, quarter after quarter. Twenty years after Bluetooth launched as a communications standard, it has finally stopped being the thing every phone has but only a fraction of a percent of people actually use.

But the success of AirPods rests on Apple's custom chip and its tie to Apple's own operating system — which is exactly the opposite of what Bluetooth's creators wanted. It is no longer the "bridge connecting every electronic device" that its inventors imagined.

Bluetooth Was Once the Hope for a Single, Unified Communications Standard

You can read the ambition in the name.

In the 1990s, several giants hit on the same idea at the same time: build a short-range wireless link. Intel had an internal project called Business-RF. Ericsson had one called MC-Link. Nokia had one called Low Power RF. All three were chasing roughly the same thing, but Ericsson finished its research first.

Ericsson engineer Jaap Haartsen had worked on the project alone since 1994. Two years later he invited the Intel and Nokia engineers chasing the same idea to sit down with him.

Jim Kardach, the Intel engineer running a similar effort, was a history buff. Since he would be working with Nordic companies, he picked up a copy of The Vikings to read up on the region's past. Any account of northern European history runs through the Danish king Harald Blatand, nicknamed "Bluetooth" because he ate so many blueberries that his teeth turned blue.

Kardach proposed "Bluetooth" as the codename. "King Harald Blatand united Scandinavia in the tenth century, just as we intend to connect PCs and mobile phones with this short-range wireless technology," he told the engineers in the room.

All three companies signed off on the codename. They saw the technology as a chance to make every electronic device interoperable — not just to connect a phone to a computer, but to connect products from different companies.

Ericsson figured that letting its phones and headsets talk to each other might lift sales of both. Intel figured that linking phones to PCs might drive its PC business.

To pull it off, they needed a common RF protocol for electronics makers worldwide. Three companies were not enough, so in 1998 they brought in representatives from IBM and Toshiba to discuss rolling the technology out globally.

In May 1998, the Bluetooth Special Interest Group (SIG), the body that would govern the technology, was formally established. The first meeting's main agenda: jointly develop a standard, interoperable protocol; agree that the technology had to be free; and agree that the chips had to be cheap and easy to install.

The meeting also approved "Bluetooth" as the official name, and combined the Old Norse runes for H and B — from "Harald Bluetooth" — to make the Bluetooth logo.

A single standard set the industry alight. By 1998, more than 200 companies had announced support for Bluetooth — even though version 1.0 had not yet shipped. Back then, if you bought a phone in China, there was about an 80 percent chance it was a Motorola, a Nokia, or an Ericsson — all backers of the standard. Microsoft and Intel, the two most important players in PC hardware and software, were on board too.

In 1999, the SIG published version 1.0 of the Bluetooth specification.

In 2000, Ericsson launched the R520, the world's first GPRS (General Packet Radio Service) phone with Bluetooth. Finland's Nokia, Britain's Psion, and Japan's TDK, among other leading vendors of the day, also released Bluetooth hardware. That year, roughly 800,000 Bluetooth devices shipped worldwide.

A year later, the hype around Bluetooth had drawn more than a billion euros in venture money. Going from standardization to commercialization to private-market investors had taken just three years.

In 2001, most of the money still came from the five founding SIG members, but plenty of small startups poured in hundreds of millions of euros too — especially the designers and makers of Bluetooth chipsets and base stations. Many were American firms, such as Silicon Wave, Conexant, and Tality; some were in the Far East, such as Japan's Alps Electric; and some were European, including BrightCom, Commil, and Britain's Cambridge Silicon Radio (CSR).

In 2018, CSR co-founder Glen Collinson recalled: "Bluetooth may be one of the most rapidly adopted standards in the world."

In radio, getting from a published spec to a shipping market usually takes years. Driven by demand, everything moved faster.

Cheap chips made mass adoption possible. While wireless chips ran above $30, CSR's Bluecore chip cost just $5 — matching Ericsson's original dream of a $5 Bluetooth chip. Within a few years, CSR had taken 60 percent of the global market.

By 2001, the Netherlands' Philips Semiconductors, Germany's Infineon, and America's Texas Instruments and National Semiconductor had all built their own Bluetooth-chip capability. PC cards looked like another important sector; that year, at least 20 companies launched products in it, including 3Com and Motorola.

The ideas came pouring in. A Danish company called BlueTags built Bluetooth luggage tags and hotel keys. Britain's Red-M built a system that let doctors reconcile hospital accounts against patients' health insurance. America's ZebraPass, backed by Nokia, built a Bluetooth wireless ticketing system for sports venues.

The vision was that any device could talk wirelessly to any other device nearby. Most market analysts were bullish. Frost & Sullivan estimated that Bluetooth shipments would grow from 11 million in 2001 to around 600 million in 2006 — about 788 million euros in revenue.

Gartner analyst Ken Dulaney warned in 2001 that inflated expectations could hurt Bluetooth: "Bluetooth has already passed the peak of inflated expectations."

Even Jaap Haartsen began to worry that hopes had been set too high.

"People are excited about Bluetooth's potential and its possibilities, but many of those expectations cannot be met in the near term," Haartsen said. "Investment in new things may go a long time without a return, and with the economy slowing, much of it is only the first step of a long journey."

Twenty Years On, Bluetooth Has Delivered Only a Fraction of What Was Imagined

Twenty years later, Bluetooth is no longer a feature phone makers tout — it comes standard on almost every phone. Like Wi-Fi, it is built in by default; on any given handset the same supplier usually packages the two together. Today, only a basic feature phone might lack it.

Bluetooth went through a tangled run from version 1.0 to 5.0, yet it never became an indispensable communications standard. It mostly connects headsets, speakers, car stereos, and mice and keyboards.

The warning signs came early.

In December 2000, SIG member 3Com gave a demo at a Bluetooth conference in Silicon Valley, trying to connect two Bluetooth laptops. After a long struggle, they could not get them to pair. A year later, at CeBIT in Hanover, German tech firm Lesswire set up more than 100 Bluetooth base stations on the show floor — the largest Bluetooth network in the world at the time — and invited attendees to try wireless transfers. The network kept reading as unavailable, and staff fiddled with it for about ten minutes in front of everyone before it finally connected.

Lesswire's project manager blamed network congestion one moment and blamed visitors for not having the right software installed the next. After the show, an Information Age reporter wrote that enthusiasm for Bluetooth was surging — and "what they fear most now is that Bluetooth's original developers were right all along: that Bluetooth is only ever a cable-replacement tool for data transfer and contact syncing."

Reality turned out worse than that early "cable replacement" framing. Bluetooth replaced only the audio cable, the mouse cable, and the keyboard cable. Even for sending data between a phone and a computer, internet services took over completely — any file of real size still goes over a USB cable or Wi-Fi.

How Bluetooth changed from version 1.0 to 5.0

1999

Version 1.0 released.

Used the 2.4 GHz band. Several manufacturers reported that their devices would not work with one another. Audio was compressed for transmission, so sound quality was poor. The protocol could not anonymize communication, creating a risk of data leakage.

2001

Version 1.1 released.

Formally folded into the IEEE 802.15.1 standard, with a transfer rate of 0.7 Mbps. But as an early design, it was prone to interference from other products on the same frequency, which degraded communication quality.

2003

Version 1.2 released.

Improved anonymization to protect users from identity-sniffing attacks and tracking, while staying backward-compatible with version 1.1.

2004

Version 2.0 released.

Introduced Enhanced Data Rate (EDR) for faster transfers. EDR's nominal rate was 3 Mbit/s, with a real-world throughput of 2.1 Mbit/s. Added duplex mode, allowing a voice call and a document or high-quality image transfer at the same time. EDR also cut power use by lowering the duty cycle.

2007

Version 2.1 released.

Stretched the interval between the mutual-confirmation signals devices exchange from about 0.1 seconds to about 0.5 seconds, easing the load on Bluetooth chips. Improved pairing and strengthened security.

2009

Version 3.0 released.

Added optional High Speed, with theoretical rates up to 24 Mbit/s — eight times Bluetooth 2.0, though Wi-Fi at the time could reach 250 Mbit/s. Introduced Enhanced Power Control (EPC) and, combined with 802.11, lowered actual idle power use.

2010

Version 4.0 released.

The first comprehensive Bluetooth protocol specification, merging three specifications into one. The biggest change was Bluetooth Low Energy (BLE), which cut power use by 90 percent versus earlier versions.

2016

Version 5.0 released.

Faster and longer-range in low-power mode: twice the transfer rate of Bluetooth 4.2 (up to 2 Mbps), four times the effective range (up to 300 meters in theory), and eight times the packet capacity. Apple's HomePod, and the iPhone 8, 8 Plus, and X, already used this standard.

Even in accessories, Bluetooth failed to unify the standard.

Buy a smart-home product — a light bulb or an air purifier — and whether it is an expensive Philips or a cheap Xiaomi, you are very likely to run into Zigbee. Rendered in Chinese as "purple bee" (紫蜂), Zigbee is a low-speed, short-range wireless networking protocol used mostly in home automation and industrial telemetry and remote control.

Use professional fitness gear and sports peripherals and you run into ANT wireless networking. The sensors that record a rider's cadence at the Tour de France, and the heart-rate straps common on gym treadmills, almost all use it. It lives mostly in health, sports, and medicine.

Premium gaming-mouse brands such as Logitech and Razer keep their own proprietary standards too. Logitech uses its own 2.4 GHz wireless format, which is faster and more precise than Bluetooth.

Designing One Standard to Make Hundreds of Hardware Companies Work Together Is Always a Nightmare

For years, Bluetooth's most immediate problem was that it was too hard to use. Nokia and Ericsson had built Bluetooth file transfer into their phones more than a decade ago, and it was once seen as a more efficient way to communicate — well suited to an era when carrier data was expensive and a single multimedia message cost a yuan.

But that kind of service required Bluetooth to work across different phones (and even PCs), across different operating systems, and across different software. Getting that many hardware makers to support each other was a nightmare.

Take a simple case: send a file from a Nokia Symbian phone and it shows up as a text message on another Symbian phone. But if the recipient used a Microsoft or Palm device — or even a different maker's Symbian phone — the transfer was all but impossible.

This was nothing like sending a photo over WeChat on Wi-Fi or 4G, where both phones reach the internet through a base station or a Wi-Fi hotspot and then trade data — the two phones never connect directly, and brand and operating system do not matter.

Soon even people inside the Bluetooth camp lost faith. Microsoft wireless expert Dr. Michael Foley said in 2001: "True interoperability between every device and every application will never arrive."

IDC analyst Chris Kozup was just as gloomy: "There are so many member companies in the SIG, and each one sees the situation from a slightly different angle. The result is a mess."

That turned out to be exactly right. The SIG set standards for Bluetooth to guarantee interoperability, but enforcement was left to many different Bluetooth Qualification Bodies (BQBs) around the world. That gave manufacturers room to improvise.

In 2001, Philips Semiconductors NV's general manager for connected products, Gerhard Heider, admitted that, in order to configure new Bluetooth features at brands' request, some BQBs had gone beyond the scope of benchmark testing.

Cutting those problems cost money. Winning full SIG support costs a company with annual revenue under US$100 million $7,500 a year in membership fees plus a $4,000 listing fee (to put the Bluetooth logo on the box) — and that excludes the many other certifications, each of which can run at least a few thousand dollars more. So many accessory makers skip the full process.

By comparison, the ANT+ standard common in fitness gear charges just $0.08 per device — which is why you can buy a cycling sensor for a few dozen yuan.

So the compatibility problem is still unsolved today, even for something as simple as a Bluetooth headset or keyboard. In 2016, someone on Hacker News — a haunt of engineers and tech enthusiasts — asked: why is Bluetooth still so terrible after all these years? The thread filled with users recounting their fights with Bluetooth devices, and engineers recounting the awful software and the chaos of standards and certifications they had run into.

More Important Still: Hardware Companies Keep Dreaming Up Features Nobody Wants — and It Keeps Being Proven

Bluetooth's trouble is not just whether it works, but what it is for. Most of the apps built around Bluetooth in those early years died quietly. Plenty of millennials have never used a piece of Bluetooth software.

In the software era, every communications technology led by hardware companies ran into similar trouble.

The most successful communications technology today is, without question, cellular data on phones — 3G and 4G. They are on nearly every smartphone now, but back when carriers decided what the mobile internet looked like, people did not even want 3G.

Carriers had already built a full suite of internet services in the 2G era. Many services we use routinely today were, in fact, available from carriers back then.

There was picture messaging via MMS; news via mobile newspapers; mobile payments charged to your phone bill (though only for games and ringtones); email via BlackBerry and carriers' BIS and BES partnerships; browsing via WAP pages; and an app store in the form of the mobile toolkit. Compared with today's internet services, carriers had it all — but nobody wanted to use any of it.

In short: usable, but not good. That is why every phone company back then had to answer to carriers, tweaking handset designs and phone software to their demands — even the biggest, Nokia, was no exception.

When 3G arrived, carriers' thinking still had not moved past the 2G era. They had spent enormous sums fighting for licenses, yet consumers had no reason to use 3G. And so, for seven full years after the network upgrade, 3G barely made a ripple. By 2006, the whole industry was asking whether the money had gone to the wrong place.

3G's biggest edge over 2G was faster downloads — but only if phones held good software and content to make users want to download in the first place. Carriers could not deliver that. Google and Apple could.

In 2008, the App Store that shipped with the iPhone 3G gave phone users their first way to browse and download apps through a channel outside the carrier — and it was a complete ecosystem. Developers built apps; users downloaded and paid for them; Apple and developers split the revenue 70/30. Carriers had no part in that loop.

Only after the iPhone 3G did Nokia's Ovi Store (later the Nokia Store), Microsoft's Windows Phone Store, and BlackBerry App World appear, one after another.

But the most successful remained Android's Google Play and the iPhone App Store. Through these two companies' operating systems and software services, consumers finally found a reason to use fast networks — and Wi-Fi, 3G, and 4G grew naturally from there.

In this story you will notice that carriers and hardware makers had plenty of time and money to set the terms for the new communications standard — and they did not.

NFC, the near-field technology once imagined as the answer to mobile payments, still has not caught on. Instead, the seemingly low-tech QR code became the default for most people in China.

Bluetooth — born under Nokia and Ericsson — was just as hemmed in by hardware makers' imagination. It has yet to produce anything genuinely good, so consumers have never found a compelling reason to use it.

Bluetooth Has Gotten More Usable in the Past Two Years — Because the Companies That Call the Shots Recast It as a Simple Accessory-Connection Technology

In 2010, Apple's iPhone 4 was among the first phones to adopt Bluetooth Low Energy (BLE), which gradually became the industry standard. The fragmented landscape eased a little.

This coincided with a broader shakeout in the phone business. The new power brokers — Apple and Google — shifted Bluetooth's role from universal communications standard to a way to connect accessories, nothing more.

The iPhone has never supported Bluetooth file transfer to other devices. Android is more open, but even Google's native Android offers no such feature. Not that many people complain. Data is cheap enough now, Wi-Fi is faster, and social networks are mature — there is no real reason to use Bluetooth for file transfer.

For accessories, manufacturers started adding their own chips to make Bluetooth's rough experience tolerable.

AirPods are the clearest example. Apple's custom W1 chip lifted AirPods above every other Bluetooth headset, creating a market where "anything that isn't AirPods is second-rate." Samsung did not design its own chip, but its chip partner Broadcom built a custom one too to keep the two earbuds in sync — plain Bluetooth was not enough.

The irony runs deeper: Bluetooth has finally produced a phenomenon in AirPods, yet AirPods cut against the technology's original purpose.

You can use any standard Bluetooth headset with an iPhone, and you can use AirPods as an ordinary Bluetooth headset with an Android phone or even a game console. But only the iPhone-plus-AirPods pairing gives you the full feature set and the best experience. The best product is not the "connect everything" device Bluetooth's creators imagined — it is the product of a closed ecosystem.

Following Apple's lead, "connect when you need to, but do not connect everything" has become big companies' attitude toward Bluetooth. All they want is for their own products to recognize one another, and they want that gap to act as a fence that keeps customers in.

Take Google's Pixel Buds. You can connect them to other devices to play music, but Google Assistant and real-time translation work only when they are paired with a Pixel phone. Paired with an iPhone, Pixel Buds are just "an ordinary Bluetooth headset."

As for Nintendo's wildly popular Switch: it has Bluetooth built in to connect controllers — but it cannot connect Bluetooth headsets.

Now Another Hardware-Led Technology Has Arrived. Will It End Differently?

5G is on the way. From carriers to chip companies, from phone makers to automakers, the mood is electric — just as Bluetooth once stirred companies worldwide.

The whole world is investing in 5G. China's three big carriers may collectively put as much as 40 billion yuan into 5G this year alone. Huawei spent 5 billion yuan on 5G R&D in 2018, winning 25 base-station contracts and shipping more than 10,000 5G base stations. The mainstream smartphone makers — Samsung, OPPO, Vivo, and Xiaomi — have all announced 5G launch timelines; in 2019, 5G will be the central selling point of new handsets.

But what exactly will consumers get from 5G?

By the time 3G arrived, social networks had already risen; video streaming sites were already common before 4G. What 5G is for remains unclear — neither China's nor America's tech giants have offered a convincing answer.

5G is faster than 4G, not just in download speed but in latency. By the GSMA's published use cases, there are four applications 4G cannot support and 5G can: autonomous driving, augmented reality, virtual reality, and the tactile internet.

AR and VR are technologies nobody uses today — content and hardware keep falling short of what consumers expect, and connectivity is the least of their problems. For autonomous driving, the bigger factor is clearly the vehicle's own computing power.

As for the oddly named tactile internet: this past March, the People's Liberation Army General Hospital, China Mobile, and Huawei ran a remote human surgery over 5G. A doctor in Beijing used the 5G network to control a brain pacemaker and operate on a Parkinson's patient at a hospital in Hainan.

It looked like a remarkably forward-looking experiment. But it rested on a lot of preconditions: no reliable Wi-Fi available; a sophisticated surgical robot on the other end; a patient who cannot move; a perfectly stable 5G signal…

A seemingly brilliant idea, built on a long list of assumptions. Just like Bluetooth in its day. Drop all those assumptions into the real world and it gets a lot harder to pull off.

5G's faster speeds and lower latency are surely the direction things are heading. But when — and how — that will actually change daily use is another question.

And the answer almost certainly will not come from telecom carriers, or from phone companies whose best idea is "downloading a movie in a few seconds."

Cover image sources: Suganth on Unsplash and GMax Studios on Unsplash