I have been following recent events in the smart lock space, particularly in the wake of Sesame (London) stopping their Indiegogo campaign, after Sesame (US) launched theirs a few days earlier on Kickstarter. As a disclaimer, I was briefly involved with Sesame (London) in mid-2014, during which I advised them to implement features and make changes that have since become evident, and implemented by almost all the other smart lock manufacturers – but not Sesame (London). It’s not very often that two genuine, competing products with the same name are launched on crowdfunding sites within days of each other.
Sesame (London), which I’ll refer to as SL from now on for clarity, sprung from a project between Marcus Kjellson and Morten Lund, who had been working on a concept of a deadbolt smart lock for almost two years. It was then renamed Sesame in mid-2014, and launched early 2015 on Indiegogo as a way to gather $75.000 to be used for an initial manufacturing run, as well as finish development. By the time I got involved with SL, the focus had become almost exclusively on aesthetic design, involving big-name architects and designers, and cool materials such as porcelain and bronze. The lock was surrounded by a marketing and design cloud, and in my view, basic functionality and lessons learned from the first wave of smart locks went ignored. I will also refer to Sesame (US) as SU from here onwards.
The first wave
I won’t go into a lot of detail on the first lot of smart locks that started appearing a couple of years ago, as Schuyler Towne has done an excellent job of it already. I will focus on the main gripes as reported by users, and how these are solved, or not, by the “Sesames”.
Smart locks and battery life
A smart lock must, as its essential feature, turn an existing deadbolt (well, any kind, but deadbolts are the easiest to support mechanically) lock from closed to open and vice-versa, upon user command or a schedule. In order to do this, an electric motor with enough torque must be used. In addition, a microprocessor, sensors and RF stack is required to control the device. For example, detecting the position of the tumbler, or implementing the “secret knock” sequence of SU’s lock requires an accelerometer. The RF stack can be Bluetooth LE and optionally Wi-Fi, if external connectivity to the internet is desired. Most first-generation smart locks that could be remote-controlled from the internet featured built-in Wi-Fi chipsets.
The main issue with Wi-Fi is that if it is to be left on 100% of the time, waiting for remote commands, the drain on the battery is huge. The motor only turns every so often, and the microprocessor and sensors can be optimized and feature “sleep” modes, but Wi-Fi must stay on. If you need to save battery, you need to turn it on periodically, or based on events such as a door shake detected by the accelerometer. If you schedule periodic connections, a remote lock or unlock is not instantaneous, voiding some of the benefits of the usual claim of “let people in remotely”.
In comparison, Bluetooth LE uses 49μA at 3V, whereas Wi-Fi uses around 100mA at 1.8V. For this simple reason, my advice to SL was to move Wi-Fi out of the lock completely, into a “hub” that plugs into AC power, and can thus be connected 24/7 to the internet – this was way before any other manufacturer had even announced plans for hubs for exactly this purpose. Any remote commands are instantly sent to the lock via Bluetooth LE, and features such as remote firmware updates are also viable.
SL attempted to fix the battery life issue by implementing a “specially designed lithium battery” which they hoped would last for a year. With Wi-Fi built into the lock, this was unlikely to happen, as has been seen in other locks. The fact that the battery is custom-made, and thus must be ordered from the manufacturer of the lock, is another barrier to adoption. SU fixed this by using lithium CR123A batteries, which is universally available in stores. A rechargeable version of the CR123A also exists, for those not wanting to spend so much on replacements. In addition, SU went straight for a Bluetooth-only lock, with a Wi-Fi hub as an option. August, Lockitron and others have all gone for Wi-Fi hubs too.
Smart locks and the desired features and functionality
The bottom line of smart locks is clear – make access easier and more convenient than using traditional keys alone. I have covered the remote lock/unlock over Wi-Fi, which requires a hub in order to make battery usage acceptable on the lock itself. The other feature touted by smart locks is automatic locking and unlocking when the user gets to his house, or leaves.
This sounds simple on the surface – how hard can it be to unlock the door if it’s locked, when the phone approaches and is detected by the lock? How hard can it be to do it in reverse, when the user leaves? Consider this “user flow” using traditional keys, of an owner arriving and entering his home:
- Get keys out of pocket
- Insert key and unlock bolt
- Take key out and open door
- Go through door
- Close door
- Throw the indoor latch on the bolt
Items 1 to 3 take place with the user outside the door, and 5-6 take place with the user inside the door. A smart lock promises to do this automagically (no pun intended). How can a smart lock know if the user is inside or outside, in order to perform the above actions in order? Location services on a phone are not accurate enough, won’t work inside buildings. Bluetooth doesn’t provide location information, so all that is left is either “secret knocks” or attempting to locate the phone by RSSI – received signal strength, as detected by the Bluetooth chipset. The issue with using RSSI is that the lock will detect the same level with the phone on particular positions relative to the door, with the phone inside and outside. Thus, the lock cannot really know if a level of -70dBm, for example, corresponds to inside or outside, as some locations on both sides of the door will result in -70dBm signal strength.
Some manufacturers have attempted to use two antennas, pointing in different directions, so that the lock can tell from the difference in signal received by the two antennas, a rough location of the phone. While this method can work, it requires calibration once the lock is installed, to account for the physical environment around it. Even then, many users report issues where locks fail to open, or fail to close once the user gets inside the house, or fail to close after leaving the house. With a Wi-Fi hub, that also contains a Bluetooth LE chipset, more accurate location can be performed, as the physical distance between antennas is larger, and the hub will definitely be on the inside of the house. Thus, if signal detected by the hub is equal or higher than that detected by the lock, the user is very likely inside the house. A simple, one-step calibration can establish the baseline of RSSI detected by lock and hub with the user standing just outside the door. It appears that SU is not going for detection of user being inside or outside at all, rather, once the phone is detected nearby, the user can use a “secret knock” to unlock. From their material, I’m assuming the lock closes once a particular vibration or acceleration pattern is detected, assumed to be the door closing shut.
Trademarks and marketing
It appears SU trademarked Sesame, and claims they started using it before SL. This will be an interesting battle, assuming SL comes back alive at all, after their disastrous start on Indiegogo. SU have played low key, but I can assume some lawyerly letters have flown around already.
Why has SL failed, while SU succeeded? In my view, due to a combination of features and pure marketing, with price also a big issue. The message “Sesame replaces your keys with your phone in seconds. It also fits on your lock in seconds, without removing or swapping anything” is easier to understand than “Sesame Smart Lock – Turn your phone into your key! Priced under $199 for a limited time!”. There is also a town market shouting style in SL’s headline, while the “under $199” feels like they are trying to fire-sale before they have even started. Not many people are in the market for a lock costing $250 upwards, with “premium” editions costing $350, plus the cost of the custom batteries added on. The messaging on SL’s campaign focuses on design and aesthetics, rather than functionality, which is what users are clamoring for after the first wave of locks hit the streets. SU appeal to the larger public, with a simple, no-frills device which you basically stick on the door with double-sided tape, and which claims to fix many of the problems with first-generation locks. There is also some appeal to the more hackery types, with references to being open-source an Arduino-compatible.
A final note on the product videos – SL uses very classy, expensive production, but without any human presence at all. SU focuses clearly on the user, and demos the whole journey in their video, using a down-to-earth guy instead of a model. The human appeal of SU is so much greater, compared to the sterile introduction of SL’s video. The videos where humans are present in SL’s campaign, feature the CEO, also in very mannered, professional tone, and giving almost no useful information that is not inferred in the page’s blurb.
Shipping dates also make SL’s campaign suffer – for such powerful claims, being able to deliver towards October 2015, means they don’t really have the internal design fully finished, or a manufacturing process in place. They have been wasting money on design and cool videos, rather than putting together a 3D-printed model and showing how it actually works, as SU have done with their prototype. SU claims they will ship May 2015, 5 months before SL.
I hate to see SL go they way they have, as I did try to help them out a few months ago. I honestly hope they resolve the trademark and whatever other issues they have, and end up launching their lock – the design is definitely cool. However, SU showed, once again, that cheap, cheerful and easy trumps visuals, design, and pure marketing. They deserve the funding and attention, if only for genuinely trying to solve the issues seen in the first generation of smart locks.