Hello friends, and welcome back to Week in Review.
Last week, we dove into the truly bizarre machinations of the NFT market. This week, we’re talking about something that’s a little bit more impactful on the current state of the web — Apple’s NeuralHash kerfuffle.
In the past month, Apple did something it generally has done an exceptional job avoiding — the company made what seemed to be an entirely unforced error.
In early August — seemingly out of nowhere** — the company announced that by the end of the year they would be rolling out a technology called NeuralHash that actively scanned the libraries of all iCloud Photos users, seeking out image hashes that matched known images of child sexual abuse material (CSAM). For obvious reasons, the on-device scanning could not be opted out of.
This announcement was not coordinated with other major consumer tech giants, Apple pushed forward on the announcement alone.
Researchers and advocacy groups had almost unilaterally negative feedback for the effort, raising concerns that this could create new abuse channels for actors like governments to detect on-device information that they regarded as objectionable. As my colleague Zach noted in a recent story, “The Electronic Frontier Foundation said this week it had amassed more than 25,000 signatures from consumers. On top of that, close to 100 policy and rights groups, including the American Civil Liberties Union, also called on Apple to abandon plans to roll out the technology.”
(The announcement also reportedly generated some controversy inside of Apple.)
The issue — of course — wasn’t that Apple was looking at find ways that prevented the proliferation of CSAM while making as few device security concessions as possible. The issue was that Apple was unilaterally making a massive choice that would affect billions of customers (while likely pushing competitors towards similar solutions), and was doing so without external public input about possible ramifications or necessary safeguards.
A long story short, over the past month researchers discovered Apple’s NeuralHash wasn’t as air tight as hoped and the company announced Friday that it was delaying the rollout “to take additional time over the coming months to collect input and make improvements before releasing these critically important child safety features.”
Having spent several years in the tech media, I will say that the only reason to release news on a Friday morning ahead of a long weekend is to ensure that the announcement is read and seen by as few people as possible, and it’s clear why they’d want that. It’s a major embarrassment for Apple, and as with any delayed rollout like this, it’s a sign that their internal teams weren’t adequately prepared and lacked the ideological diversity to gauge the scope of the issue that they were tackling. This isn’t really a dig at Apple’s team building this so much as it’s a dig on Apple trying to solve a problem like this inside the Apple Park vacuum while adhering to its annual iOS release schedule.
Image Credits: Bryce Durbin / TechCrunch /
Apple is increasingly looking to make privacy a key selling point for the iOS ecosystem, and as a result of this productization, has pushed development of privacy-centric features towards the same secrecy its surface-level design changes command. In June, Apple announced iCloud+ and raised some eyebrows when they shared that certain new privacy-centric features would only be available to iPhone users who paid for additional subscription services.
You obviously can’t tap public opinion for every product update, but perhaps wide-ranging and trail-blazing security and privacy features should be treated a bit differently than the average product update. Apple’s lack of engagement with research and advocacy groups on NeuralHash was pretty egregious and certainly raises some questions about whether the company fully respects how the choices they make for iOS affect the broader internet.
Delaying the feature’s rollout is a good thing, but let’s all hope they take that time to reflect more broadly as well.
** Though the announcement was a surprise to many, Apple’s development of this feature wasn’t coming completely out of nowhere. Those at the top of Apple likely felt that the winds of global tech regulation might be shifting towards outright bans of some methods of encryption in some of its biggest markets.
Back in October of 2020, then United States AG Bill Barr joined representatives from the UK, New Zealand, Australia, Canada, India and Japan in signing a letter raising major concerns about how implementations of encryption tech posed “significant challenges to public safety, including to highly vulnerable members of our societies like sexually exploited children.” The letter effectively called on tech industry companies to get creative in how they tackled this problem.
Here are the TechCrunch news stories that especially caught my eye this week:
LinkedIn kills Stories
You may be shocked to hear that LinkedIn even had a Stories-like product on their platform, but if you did already know that they were testing Stories, you likely won’t be so surprised to hear that the test didn’t pan out too well. The company announced this week that they’ll be suspending the feature at the end of the month. RIP.
FAA grounds Virgin Galactic over questions about Branson flight
While all appeared to go swimmingly for Richard Branson’s trip to space last month, the FAA has some questions regarding why the flight seemed to unexpectedly veer so far off the cleared route. The FAA is preventing the company from further launches until they find out what the deal is.
Apple buys a classical music streaming service
While Spotify makes news every month or two for spending a massive amount acquiring a popular podcast, Apple seems to have eyes on a different market for Apple Music, announcing this week that they’re bringing the classical music streaming service Primephonic onto the Apple Music team.
TikTok parent company buys a VR startup
It isn’t a huge secret that ByteDance and Facebook have been trying to copy each other’s success at times, but many probably weren’t expecting TikTok’s parent company to wander into the virtual reality game. The Chinese company bought the startup Pico which makes consumer VR headsets for China and enterprise VR products for North American customers.
Twitter tests an anti-abuse ‘Safety Mode’
The same features that make Twitter an incredibly cool product for some users can also make the experience awful for others, a realization that Twitter has seemingly been very slow to make. Their latest solution is more individual user controls, which Twitter is testing out with a new “safety mode” which pairs algorithmic intelligence with new user inputs.
Some of my favorite reads from our Extra Crunch subscription service this week:
Our favorite startups from YC’s Demo Day, Part 1
“Y Combinator kicked off its fourth-ever virtual Demo Day today, revealing the first half of its nearly 400-company batch. The presentation, YC’s biggest yet, offers a snapshot into where innovation is heading, from not-so-simple seaweed to a Clearco for creators….”
“…Yesterday, the TechCrunch team covered the first half of this batch, as well as the startups with one-minute pitches that stood out to us. We even podcasted about it! Today, we’re doing it all over again. Here’s our full list of all startups that presented on the record today, and below, you’ll find our votes for the best Y Combinator pitches of Day Two. The ones that, as people who sift through a few hundred pitches a day, made us go ‘oh wait, what’s this?’
All the reasons why you should launch a credit card
“… if your company somehow hasn’t yet found its way to launch a debit or credit card, we have good news: It’s easier than ever to do so and there’s actual money to be made. Just know that if you do, you’ve got plenty of competition and that actual customer usage will probably depend on how sticky your service is and how valuable the rewards are that you offer to your most active users….”
On average, men and women speak roughly 15,000 words per day. We call our friends and family, log into Zoom for meetings with our colleagues, discuss our days with our loved ones, or if you’re like me, you argue with the ref about a bad call they made in the playoffs.
Hospitality, travel, IoT and the auto industry are all on the cusp of leveling-up voice assistant adoption and the monetization of voice. The global voice and speech recognition market is expected to grow at a CAGR of 17.2% from 2019 to reach $26.8 billion by 2025, according to Meticulous Research. Companies like Amazon and Apple will accelerate this growth as they leverage ambient computing capabilities, which will continue to push voice interfaces forward as a primary interface.
As voice technologies become ubiquitous, companies are turning their focus to the value of the data latent in these new channels. Microsoft’s recent acquisition of Nuance is not just about achieving better NLP or voice assistant technology, it’s also about the trove of healthcare data that the conversational AI has collected.
Our voice technologies have not been engineered to confront the messiness of the real world or the cacophony of our actual lives.
Google has monetized every click of your mouse, and the same thing is now happening with voice. Advertisers have found that speak-through conversion rates are higher than click-through conversation rates. Brands need to begin developing voice strategies to reach customers — or risk being left behind.
Voice tech adoption was already on the rise, but with most of the world under lockdown protocol during the COVID-19 pandemic, adoption is set to skyrocket. Nearly 40% of internet users in the U.S. use smart speakers at least monthly in 2020, according to Insider Intelligence.
Yet, there are several fundamental technology barriers keeping us from reaching the full potential of the technology.
By the end of 2020, worldwide shipments of wearable devices rose 27.2% to 153.5 million from a year earlier, but despite all the progress made in voice technologies and their integration in a plethora of end-user devices, they are still largely limited to simple tasks. That is finally starting to change as consumers demand more from these interactions, and voice becomes a more essential interface.
In 2018, in-car shoppers spent $230 billion to order food, coffee, groceries or items to pick up at a store. The auto industry is one of the earliest adopters of voice AI, but in order to really capture voice technology’s true potential, it needs to become a more seamless, truly hands-free experience. Ambient car noise still muddies the signal enough that it keeps users tethered to using their phones.
Getting inside the mind of customers is a challenge as behaviors and demands shift, but Clootrack believes it has cracked the code in helping brands figure out how to do that.
It announced $4 million in Series A funding, led by Inventus Capital India, and included existing investors Unicorn India Ventures, IAN Fund and Salamander Excubator Angel Fund, as well as individual investment from Jiffy.ai CEO Babu Sivadasan. In total, the company raised $4.6 million, co-founder Shameel Abdulla told TechCrunch.
Clootrack is a real-time customer experience analytics platform that helps brands understand why customers stay or churn. Shameel Abdulla and Subbakrishna Rao, who both come from IT backgrounds, founded the company in 2017 after meeting years prior at Jiffstore, Abdulla’s second company that was acquired in 2015.
Clootrack team. Image Credits: Clootrack
Business-to-consumer and consumer brands often use customer satisfaction metrics like Net Promoter Score to understand the customer experience, but Abdulla said current methods don’t provide the “why” of those experiences and are slow, expensive and error-prone.
“The number of channels has increased, which means customers are talking to you, expressing their feedback and what they think in multiple places,” he added. “Word of mouth has gone digital, and you basically have to master the art of selling online.”
Clootrack turns the customer experience data from all of those first-party and third-party touchpoints — website feedback, chat bots, etc. — into granular, qualitative insights that give brands a look at drivers of the experience in hours rather than months so that they can stay on top of fast-moving trends.
Abdulla points to data that show a customer’s biggest driver of brand switch is the experience they receive. And, that if brands can reduce churns by 5%, they could be looking at an increase in profits of between 25% and 95%.
Most of the new funding will go to product development so that all data aggregations are gathered from all possible touchpoints. His ultimate goal is to be “the single platform for B2C firms.”
The company is currently working with over 150 customers in the areas of retail, direct-to-consumer, banking, automotive, travel and mobile app-based services. It is growing nine times year over year in revenue. It is mainly operating in India, but Clootrack is also onboarding companies in the U.S. and Europe.
Parag Dhol, managing director of Inventus, said he has known Abdulla for over five years. He had looked at one of Abdulla’s companies for investment, but had decided against it due to his firm being a Series A investor.
Dhol said market research needs an overhaul in India, where this type of technology is lagging behind the U.S.
“Clootrack has a very complementary team with Shameel being a complete CEO in terms of being a sales guy and serial entrepreneur who has learned his lessons, and Subbu, who is good at technology,” he added. “As CMOs realize the value in their unstructured data inside of their own database of the customer reviews and move to real-time feedback, these guys could make a serious dent in the space.”
Cohort analysis is a way of evaluating your business that involves grouping customers into “cohorts” and observing how they behave over time. A commonly used approach is monthly cohort analysis, where customers are grouped by the month they signed up, allowing you to observe how someone who joined in November compares to someone who signed up the month before.
Cohort analysis gives you a multivariable, forward-looking view of your business compared to more simple and static values like averages or totals.
Cohort analysis is flexible and can be used to analyze a variety of performance metrics including revenue, acquisition costs and churn.
Let’s imagine you’re the CMO of the “Bluetooth Coffee Company.” You sell a tech-enabled “coffee composer” that brews coffee, tracks consumption and orders replacement coffee when users are running low. The longer your customers are subscribers, the more money you make. You recently ran a Black Friday feature on a popular deals site and you’re interested to know if you should run it again.
The chart below is a simple analysis you might do to gauge your marketing performance. It shows the total customers added each month, and a clear spike in November following the Black Friday promotion. At first glance, things look good — you brought in more than double the monthly customers in November compared to October.
Image Credits: Sagard & Portage Ventures
But before you rebook the promotion, you should ask if these new Black Friday consumers are as valuable as they seem. Comparing monthly customer percentage is a good way to find out.
Below is a monthly cohort analysis of new customers between September 2020 and February 2021. Like our previous chart, we’ve listed the monthly cohort size, but we’ve also included the customer engagement rate (calculated by dividing daily active users by monthly active users or DAU/MAU for each month (M1 is month 1, M2 is month 2, and so on).
This analysis lets us see how the customer engagement of each monthly cohort compares to the next.
Image Credits: Sagard & Portage Ventures
From the figures above, we see that most cohorts have a customer engagement rate in their first month (M1, 42%-46%), meaning 42%-46% of new customers use the coffee composer everyday. The November cohort however has materially lower engagement (M1, 30%), and remains lower in subsequent months (M2, 26%) and (M3, 27%). Interestingly, the customer engagement rate only drops with the November cohort, returning to normal with the December cohort (M1, 45%).
If the past 18 months is any indication, the nature of the workplace is changing. And while Box and Zoom already have integrations together, it makes sense for them to continue to work more closely.
Their newest collaboration is the Box app for Zoom, a new type of in-product integration that allows users to bring apps into a Zoom meeting to provide the full Box experience.
While in Zoom, users can securely and directly access Box to browse, preview and share files from Zoom — even if they are not taking part in an active meeting. This new feature follows a Zoom integration Box launched last year with its “Recommended Apps” section that enables access to Zoom from Box so that workflows aren’t disrupted.
The companies’ chief product officers, Diego Dugatkin with Box and Oded Gal with Zoom, discussed with TechCrunch why seamless partnerships like these are a solution for the changing workplace.
With digitization happening everywhere, an integration of “best-in-breed” products for collaboration is essential, Dugatkin said. Not only that, people don’t want to be moving from app to app, instead wanting to stay in one environment.
“It’s access to content while never having to leave the Zoom platform,” he added.
It’s also access to content and contacts in different situations. When everyone was in an office, meeting at a moment’s notice internally was not a challenge. Now, more people are understanding the value of flexibility, and both Gal and Dugatkin expect that spending some time at home and some time in the office will not change anytime soon.
As a result, across the spectrum of a company, there is an increasing need for allowing and even empowering people to work from anywhere, Dugatkin said. That then leads to a conversation about sharing documents in a secure way for companies, which this collaboration enables.
The new Box and Zoom integration enables meeting in a hybrid workplace: chat, video, audio, computers or mobile devices, and also being able to access content from all of those methods, Gal said.
“Companies need to be dynamic as people make the decision of how they want to work,” he added. “The digital world is providing that flexibility.”
This long-term partnership is just scratching the surface of the continuous improvement the companies have planned, Dugatkin said.
Dugatkin and Gal expect to continue offering seamless integration before, during and after meetings: utilizing Box’s cloud storage, while also offering the ability for offline communication between people so that they can keep the workflow going.
“As Diego said about digitization, we are seeing continuous collaboration enhanced with the communication aspect of meetings day in and day out,” Gal added. “Being able to connect between asynchronous and synchronous with Zoom is addressing the future of work and how it is shaping where we go in the future.”
Explosion, a company that has combined an open source machine learning library with a set of commercial developer tools, announced a $6 million Series A today on a $120 million valuation. The round was led by SignalFire, and the company reported that today’s investment represents 5% of its value.
Oana Olteanu from SignalFire will be joining the board under the terms of the deal, which includes warrants of $12 million in additional investment at the same price.
“Fundamentally, Explosion is a software company and we build developer tools for AI and machine learning and natural language processing. So our goal is to make developers more productive and more focused on their natural language processing, so basically understanding large volumes of text, and training machine learning models to help with that and automate some processes,” company co-founder and CEO Ines Montani told me.
The company started in 2016 when Montani met her co-founder, Matthew Honnibal in Berlin where he was working on the spaCy open source machine learning library. Since then, that open source project has been downloaded over 40 million times.
In 2017, they added Prodigy, a commercial product for generating data for the machine learning model. “Machine learning is code plus data, so to really get the most out of the technologies you almost always want to train your models and build custom systems because what’s really most valuable are problems that are super specific to you and your business and what you’re trying to find out, and so we saw that the area of creating training data, training these machine learning models, was something that people didn’t pay very much attention to at all,” she said.
The next step is a product called Prodigy Teams, which is a big reason the company is taking on this investment. “Prodigy Teams is [a hosted service that] adds user management and collaboration features to Prodigy, and you can run it in the cloud without compromising on what people love most about Prodigy, which is the data privacy, so no data ever needs to get seen by our servers,” she said. They do this by letting the data sit on the customer’s private cluster in a private cloud, and then use Prodigy Team’s management features in the public cloud service.
Today, they have 500 customers using Prodigy including Microsoft and Bayer in addition to the huge community of millions of open source users. They’ve built all this with just 17 people, even as they continue to slowly add employees, expecting to reach 20 by the end of the year.
She believes if you’re thinking too much about diversity in your hiring process, you probably have a problem already. “If you go into hiring and you’re thinking like, oh, how can I make sure that the way I’m hiring is diverse, I think that already shows that there’s maybe a problem,” she said.
“If you have a company, and it’s 50 dudes in their 20s, it’s not surprising that you might have problems attracting people who are not white dudes in their 20s. But in our case, our strategy is to hire good people and good people are often very diverse people, and again if you play by the [startup] playbook, you could be limited in a lot of other ways.”
She said that they have never seen themselves as a traditional startup following some conventional playbook. “We didn’t raise any investment money [until now]. We grew the team organically, and we focused on being profitable and independent [before we got outside investment],” she said.
But more than the money, Montani says that they needed to find an investor that would understand and support the open source side of the business, even while they got capital to expand all parts of the company. “Open source is a community of users, customers and employees. They are real people, and [they are not] pawns in [some] startup game, and it’s not a game. It’s real, and these are real people,” she said.
“They deserve more than just my eyeballs and grand promises. […] And so it’s very important that even if we’re selling a small stake in our company for some capital [to build our next] product [that open source remains at] the core of our company and that’s something we don’t want to compromise on,” Montani said.