Blog

Today marks the start of Facebook CEO Mark Zuckerberg’s much-anticipated trip to Washington as he attends a hearing with the Senate, before moving on to a Congressional hearing tomorrow.

Away from the U.S. political capital, Zuckerberg is engaged in serious discussions about Myanmar with a group of six civil society organizations in the country who took umbrage at his claim that Facebook’s systems had prevented messages aimed at inciting violence between Buddhists and Muslims last September.

Following an open letter to Facebook on Friday that claimed the social network had relied on local sources and remains ill-equipped to handle hate speech, Zuckerberg himself stepped in to personally respond.

“Thank you for writing it and I apologize for not being sufficiently clear about the important role that your organizations play in helping us understand and respond to Myanmar-related issues, including the September incident you referred to,” Zuckerberg wrote.

“In making my remarks, my intention was to highlight how we’re building artificial intelligence to help us better identify abusive, hateful or false content even before it is flagged by our community,” he added.

Zuckerberg also claimed Facebook is working to implement new features that include the option to report inappropriate content inside Messenger, and adding more Burmese language reviewers — two suggestions that the Myanmar-based group had raised.

The group has, however, fired back again to criticize Zuckerberg’s response which it said is “nowhere near enough to ensure that Myanmar users are provided with the same standards of care as users in the U.S. or Europe.”

Young men browse their Facebook wall on their smartphones as they sit in a street in Yangon on August 20, 2015. Facebook remains the dominant social network for US Internet users, while Twitter has failed to keep apace with rivals like Instagram and Pinterest, a study showed. AFP PHOTO / Nicolas ASFOURI (Photo credit should read NICOLAS ASFOURI/AFP/Getty Images)

In particular, the six companies are asking Facebook and Zuckerberg to give information around its efforts, including the number of abuse reports it has received, how many have been removed, how quickly it has been done, and its progress on banning accounts.

In addition, the group asked for clarity on the number of Burmese content reviewers on staff, the exact mechanisms that are in place for detecting hate speech, and an update on what action Facebook has taken following its last meeting with the group in December.

“When things go wrong in Myanmar, the consequences can be really serious — potentially disastrous,” it added.

The Cambridge Analytica story has become mainstream news in the U.S. and other parts of the world, yet less is known of Facebook’s role in spreading religious hatred in Myanmar, where the government stands accused of ethnic cleansing following its treatment of the minority Muslim Rohingya population.

A recent UN Fact-Finding Mission concluded that social media has played a “determining role” in the crisis, which Facebook the chief actor.

“We know that the ultranationalist Buddhists have their own [Facebook pages] and really [are] inciting a lot of violence and a lot of hatred against the Rohingya or other ethnic minorities. I’m afraid that Facebook has now turned into a beast, [instead of] what it was originally intended to be used [for],” the UN’s Yanghee Lee said to media.

Close to 30 million of Myanmar’s 50 million population is said to use the social network, making it a hugely effective way to reach large audiences.

“There’s this notion to many people [in Myanmar] that Facebook is the internet,” Jes Petersen, CEO of Phandeeyar — one of the companies involved in the correspondence with Zuckerberg — told TechCrunch in an interview last week.

Despite that huge popularity and high levels of abuse that Facebook itself has acknowledged, the social network does not have an office in Myanmar. In fact, its Burmese language reviewers are said to be stationed in Ireland while its policy team is located in Australia.

That doesn’t seem like the right combination, but it is also unclear whether Facebook is prepared to make changes to focus on user safety in Myanmar. The company declined to say whether it had plans to open an office on the ground when we asked last week.

Here’s Zuckerberg’s letter in full:

Dear Htaike Htaike, Jes, Victoire, Phyu Phyu and Thant,

I wanted to personally respond to your open letter. Thank you for writing it and I apologize for not being sufficiently clear about the important role that your organizations play in helping us understand and respond to Myanmar-related issues, including the September incident you referred to.

In making my remarks, my intention was to highlight how we’re building artificial intelligence to help us better identify abusive, hateful or false content even before it is flagged by our community.

These improvements in technology and tools are the kinds of solutions that your organizations have called on us to implement and we are committed to doing even more. For example, we are rolling out improvements to our reporting mechanism in Messenger to make it easier to find and simpler for people to report conversations.

In addition to improving our technology and tools, we have added dozens more Burmese language reviewers to handle reports from users across all our services. We have also increased the number of people across the company on Myanmar-related issues and we now we have a special product team working to better understand the specific local challenges and build the right tools to help keep people there safe.

There are several other improvements we have made or are making, and I have directed my teams to ensure we are doing all we can to get your feedback and keep you informed.

We are grateful for your support as we map out our ongoing work in Myanmar, and we are committed to working with you to find more ways to be responsive to these important issues.

Mark

And the group’s reply:

Dear Mark,

Thank you for responding to our letter from your personal email account. It means a lot.

We also appreciate your reiteration of the steps Facebook has taken and intends to take to improve your performance in Myanmar.

This doesn’t change our core belief that your proposed improvements are nowhere near enough to ensure that Myanmar users are provided with the same standards of care as users in the US or Europe.

When things go wrong in Myanmar, the consequences can be really serious – potentially disastrous. You have yourself publicly acknowledged the risk of the platform being abused towards real harm.

Like many discussions we have had with your policy team previously, your email focuses on inputs. We care about performance, progress and positive outcomes.

In the spirit of transparency, we would greatly appreciate if you could provide us with the following indicators, starting with the month of March 2018:

• How many reports of abuse have you received?
• What % of reported abuses did your team ultimately remove due to violations of the community standards?
• How many accounts were behind flagging the reports received?
• What was the average time it took for your review team to provide a final response to users of the reports they have raised?
• What % of the reports received took more than 48 hours to receive a review?
• Do you have a target for review times? Data from our own monitoring suggests that you might have an internal standard for review – with most reported posts being reviewed shortly after the 48 hrs mark. Is this accurate?
• How many fake accounts did you identify and removed?
• How many accounts did you subject to a temporary ban? How many did you ban from the platform?

Improved performance comes with investments and we would also like to ask for more clarifications around these. Most importantly, we would like to know:

• How many Myanmar speaking reviewers did you have, in total, as of March 2018? How many do you expect to have by the end of the year? We are specifically interested in reviewers working on the Facebook service and looking for full-time equivalents figure.
• What mechanisms do you have in place for stopping repeat offenders in Myanmar? We know for a fact that fake accounts remain a key issue and that individuals who were found to violate the community standards on a number of occasions continue to have a presence on the platform.
• What steps have you taken to date to address the duplicate posts issue we raised in the briefing we provided your team in December 2017?

We’re enclosing our December briefing for your reference, as it further elaborates on the challenges we have been trying to work through with Facebook.

Best,

Cambridge, U.K.-based startup Spectral Edge has closed a $5.3M Series A funding round from existing investors Parkwalk Advisors and IQ Capital.

The team, which in 2014 spun the business out of academic research at the University of East Anglia, has developed a mathematical technique for improving photographic imagery in real-time, also using machine learning technology. 

As we’ve reported previously, their technology — which can be embedded in software or in silicon — is designed to enhance pictures and videos on mass-market devices. Mooted use cases include for enhancing low light smartphone images, improving security camera footage or even for drone cameras. 

This month Spectral Edge announced its first customer, IT services provider NTT data, which said it would be incorporating the technology into its broadcast infrastructure offering — to offer its customers an “HDR-like experience”, via improved image quality, without the need for them to upgrade their hardware.

“We are in advanced trials with a number of global tech companies — household names — and hope to be able to announce more deals later this year,” CEO Rhodri Thomas tells us, adding that he expects 2-3 more deals in the broadcast space to follow “soon”, and enhance viewing experiences “in a variety of ways”.

On the smartphone front, Thomas says the company is waiting for consumer hardware to catch up — noting that RGB-IR sensors “haven’t yet begun to deploy on smartphones on a great scale”.

Once the smartphone hardware is there he reckons its technology will be able to help with various issues such as white balancing and bokeh processing.

“Right now there is no real solution for white balancing across the whole image [on smartphones] — so you’ll get areas of the image with excessive blues or yellows, perhaps, because the balance is out — but our tech allows this to be solved elegantly and with great results,” he suggests. “We also can support bokeh processing by eliminating artifacts that are common in these images.”

The new funding is going towards ramping up Spectral Edge’s efforts to commercialize its tech, including by growing the R&D team to 12 — with hires planned for specialists in image processing, machine learning and embedded software development.

The startup will also focus on developing real-world apps for smartphones, webcams and security applications alongside its existing products for the TV & display industries.

“The company is already very IP strong, with 10 patent families in the world (some granted, some filed and a couple about to be filed),” says Thomas. “The focus now is productizing and commercializing.”

“In a year, I expect our technology to be launched or launching on major flagship [smartphone] devices,” he adds. “We also believe that by then our CVD (color vision deficiency) product, Eyeteq, is helping millions of people suffering from color blindness to enjoy significantly better video experiences.”

Ahead of Mark Zuckerberg’s Senate testimony today, Facebook has rolled out a number of product updates — including a bounty hunting program of up to $40,000 for user data violations — meant to address (and blunt) the criticism he’s likely to face.

The bounties start at $500, according to a report by CNBC, and will be awarded if certain conditions are met.

First announced amid a slew of updates Zuckerberg offered up in March as the scandal around abuse of user data by the political consulting firm Cambridge Analytica was first coming to light, the new bounty program is modeled off of Facebook’s attempts to combat hackers with a $1 million bounty.

It’s the second product announcement today, following news that Facebook would stop apps from accessing user data if they haven’t been launched within 90 days.

To be eligible for the bounty, the offending app must impact more than 10,000 Facebook users and show a clear pattern of abuse and not “collection” (in this case, I’m assuming abuse would qualify as transferring the data to a third party without permission).

Facebook also stipulated that it should be a case that the company isn’t already actively investigating.

Examples of “out of scope” scenarios include: scraping, malware, social engineering applications, and cases involving other Facebook companies (like Instagram).

Facebook goes on to assure that if whistleblowers comply with the company’s policy, then the company won’t sue them (which is very big of Facebook). It also tries to ensure that all of the issues are kept quiet and far away from the meddling of the media which could blow the whole thing up and force company executives to testify in front of Congressional hearings.

Here are some other details from the program:

• • You give us time to investigate and act on an issue that you report before making any information about the report public or sharing such information with others.
  • You make a good faith effort to avoid privacy violations and disruptions to others, including (but not limited to) unauthorized access to or destruction of data, and interruption or degradation of our or another’s services.
  • You provide us with the Facebook data we request after we request it.
  • You do not violate any other applicable laws or regulations, including (but not limited to) laws and regulations prohibiting the unauthorized access to data. Again, do not submit any data to us that you obtained unlawfully.

Those are the requirements for folks who submit bounties to Facebook. For its part, Facebook agrees to investigate and validate the reports; determine the amount of the bounty; takes responsibility for publishing reports and updates on the bounty; and indicates that one bounty is paid per abuse (if multiple reports of abuse are submitted, it’s first come, first served for payouts).

No word yet on whether Facebook will be paying individual journalists or The Guardian or the The New York Times retroactively.

There’s a new venture fund in town from some familiar faces.

Carey Lai, who previously worked at Intel Capital and IVP, is joining forces with Paul Yeh, formerly of Kleiner Perkins.

They’re calling it Conductive Ventures and it’s launching with $100 million under management. They’ll be investing in “expansion stage” companies across enterprise software and hardware categories, meaning Series A, Series B and beyond.

Check sizes will be between $2 million and $7 million dollars. They expect to invest in 10-15 companies for this first fund.

Conductive will be looking for “early product market fit with customer success,” Lai told TechCrunch. Then the plan is to “help them grow their businesses abroad.”

It’s not a corporate venture arm, but Conductive has Panasonic as its sole LP. Because of this, there will be a special focus on helping North American startups expand into Asia, particularly Japan.

Lai and Yeh touted “connections to Foxconn” and also ties to Taiwan to help them succeed overseas.

They also said they want to be hands-on when it comes to growth. Conductive will place an emphasis on improving margins, aiming to accelerate revenue and reduce costs.

The two were roommates when they were younger and think that they will get along especially well as an investment team.

So far, they’ve made four investments. There’s Ambiq Micro, a semiconductor manufacturer; CSC Generation, for consumer leasing; Desktop Metal, in 3D printing; and Sprinklr, for customer experience management. Lai has served on the board of Sprinklr. They hope to continue to take board seats.

Not to get ahead of things, but they are already thinking about fund two. Yeh said that it will be in “a couple years” and “slightly higher, slightly bigger” in size.

Sam Kim, co-founder and CEO of Lucidity, said blockchain technology offers a solution to cut down on fraud and bring more transparency to digital advertising.

Kim was previously chief operating officer at The Mobile Majority, a mobile ad company where he said he saw the challenges of reconciling the data used by different parts of the ad supply chain. So at Lucidity, his team has created technology on the Ethereum blockchain to create smart contracts that track this data across the ecosystem.

“What we’ve built is a protocol for industry participants come to a consensus what events are valid and what are invalid,” he said. “We allow all the participants trafficking that ad to submit that data … so that basically becomes a single source of truth across the industry.”

On top of the protocol, Lucidity has created products focused on specific issues. There’s one around fee transparency, where every participant in the supply chain submits the information about the fee they charged and verification that they actually ran the impression, which Kim said could replace a drawn-out reconciliation process. And there’s another product for confirming publisher identity and flagging fake bid requests.

When I asked whether ad platforms and publishers will feel comfortable sharing this data in a new way, Kim said, “As an industry, there’s a lot of pressure from the advertisers to become more transparent.”

“The way we designed it, it will benefit those who are honest and transparent,” he added. “Those who aren’t, it’ll be harmful to them. We expect to see adoption from those who are completely transparent and honest, and hopefully drive the industry in that direction.”

As for potential privacy concerns, Kim noted that Lucidity doesn’t currently include consumer data. That could change in the future, but if it does, Kim said it would be done in a way that ensured “privacy was addressed.”

Lucidity has been in beta testing since earlier this year, and it’s already signed up partners from the ad world including Audience Group and Digital Remedy.

Kim said the goal is to get “the benefits of the blockchain — disintermediation, transparency and trust — but deployed in a way that’s not cumbersome and difficult for the current players.” So if you’re an advertiser, you don’t need to worry too much about the underlying technology; you just get a dashboard with the data you want.

At the same time, Kim said the company is using “cutting edge technology,” for example by implementing Plasma, a side-chain processing technology that’s supposed to address Ethereum’s speed and scaling issues. And it’s partnered with IAB Tech Lab’s Blockchain Working Group, the Enterprise Ethereum Association and Hyperledger.

The startup was previously known as Kr8os (from the Greek word kratos, which was one of the ingredients for democracy). Today it’s rebranding as Lucidity.

“We’d been thinking about the rebrand for some time,” Kim said. “Kr8os is a process we used to ultimately drive our vision for Lucidity.”

Facebook has rolled out a change that limits apps’ ability to access users’ data. The change comes hours before Facebook CEO Mark Zuckerberg’s testimony before Congress where it’s expected he will have to answer questions about how the company handles its users’ personal information, in wake of the Cambridge Analytica scandal which saw the personal data of 87 million Facebook users compromised.

Among many other new restrictions to its API platform announced last week, Facebook said that it would soon introduce a stricter review process for use of Facebook Login for apps, and it would block apps from pulling users’ personal data after three months of non-use.

Now, that change to apps’ ability to access user data has begun to roll out, Facebook says.

All Facebook platform apps won’t be affected immediately – the change will roll out gradually over the weeks ahead.  Between April 9 and April 21, tokens for the users who have not actively logged into a developer’s app and granted consent to permissions in the last 90 days will expire, Facebook says.

Facebook suggests that app publishers monitor their app for any issues that may occur as a result. Apps may run into issues with regard to the expired tokens, and will need to make sure they’ve designed their app to either re-prompt the user to login again with Facebook, or show an optional user interface which allows the app to refresh its access to users’ Facebook data with consent.

The apps will have to send users through the Facebook Login process every 90 days, and the person logging in has to agree to the data permissions by tapping “Continue,” Facebook explains in its announcement.

“We believe this immediate access update helps build trust and leads to stronger connections within our ecosystem,” reads the Facebook blog post.

A number of apps over the years adopted Facebook Login to offer users an easier way to sign into their own service, while giving the app maker the ability to access users’ Facebook data. Some users preferred the Facebook Login option, as it meant they didn’t have to remember so many different passwords. Others, including those who didn’t have a Facebook account (or those who perhaps rightly didn’t trust Facebook) found the practice infuriating.

Also upsetting is that there was no sort of deprecation policy in place for the apps people no longer used. That led to users being fairly shocked to discover long lists of apps they hadn’t touched in years with lingering access to their data. Facebook recently addressed this issue as well, with the introduction of a bulk app removal tool that lets users delete apps from their account entirely. 

It also announced a series of changes to how developers can use its APIs, including Instagram APIs, which largely involve locking down its platform, then figuring out which developers actually require (and deserve) any heightened access in order for their app to function.

These sorts of changes are a critical part of what Zuckerberg will have to testify to today – because it wasn’t that Facebook directly handed over 87 million users’ personal data – it had just designed a platform that let apps easily collect it without users’ knowledge or consent.

By now it’s crystal clear to just about everyone that the password is a weak and frankly meaningless form of authentication, yet most of us still live under the tyranny of the password. This, despite the fact it places a burden on the user, is easily stolen and mostly ineffective. Today, two standards bodies, FIDO and W3C announced a better way, a new password free protocol for the web called WebAuthn.

The major browser makers including Google, Mozilla and Microsoft have all agreed to incorporate the final version of the protocol, which allow websites to bypass the pesky password in favor of an external authenticator such as a security key or you mobile phone. These devices will communicate directly with the website via Bluetooth, USB or NFC. The standards body has referred to this as ‘phishing-proof’.

Yes, by switching to this method, not only will you eliminate the need for a password — or to come up with a 20-character one every few weeks to please the security gods — but the whole reason for that kind of security farce will disappear. Without passwords, we can eliminate many of the common security threats out there including phishing, man-in-the-middle attacks and general abuse of stolen credentials. That’s because using a system like this, there wouldn’t be anything to steal. The authentication token would only last as long as it takes to authenticate the user and no more and would require a specific device to authenticate.

The WebAuthn specification offers several examples of how this could work. In one example, you are working on a laptop and you access a website that requires you to log in. Instead of a user name and password, you get a prompt to check your phone. You tap the prompt on your phone and you are logged in without the need for entering anything.

Brett McDowell, executive director of the FIDO Alliance certainly saw the beauty of this new approach. “After years of increasingly severe data breaches and password credential theft, now is the time for service providers to end their dependency on vulnerable passwords and one-time-passcodes and adopt phishing-resistant FIDO Authentication for all websites and applications,” McDowell said in a statement..

WebAuthn is not quite ready for final release just yet, but it has reached the “Candidate Recommendation (CR) stage, which means it’s being recommended to the standards bodies for final approval.

No security method is fool-proof, of course, and it probably won’t take long for someone to find a hole in this approach too, but at the very least it’s a step in the right direction. It is long past time that we come up with a new password-free authentication technique and WebAuthn just might be the answer to the long-standing problem of passwords.

Landing pages might not sound like the most exciting problem for a startup to tackle, but Tyson Quick, founder and CEO of Instapage, said that they’re a real issue for marketers.

Quick was previously working on a marketing product called Jounce, where he was frustrated by the fact that even after someone had clicked on a Jounce ad and ended up on the company’s website, only 3 percent of them were converting.

“It was kind of ridiculous, since they’d already engaged with the ad,” Quick said. So just by creating a few personalized landing pages that targeted “major” customer segments, Jounce was able to increase conversions to around 20 percent.

In 2012, Quick decided to switch the company’s focus to Instapage in, and it now offers a broader platform for creating and testing different landing pages that are personalized for different visitors.

“We can be this middleware between ad tech and traditional marketing and sales tech,” Quick said. “All this technology has made it so the ad-buying side has sophisticated targeting and relevancy. And once you capture a lead, you’ve got marketing automation that pushes them into your CRM. It’s the point of conversion that’s missing.”

Instapage now has more than 15,000 customers who are seeing average conversion rates about 22 percent. And it’s announcing that it’s raised $15 million in Series A funding from Morgan Stanley Expansion Capital.

Quick admitted that it hasn’t just been a straight line to success. He recalled moving from Salt Lake City to San Francisco and living in his car for part of the month so that he could rent out his apartment on Airbnb.

“I didn’t want to get a job,” he explained — and the Airbnb money helped him devote his time to Instapage.

Quick also said it took some time to find the ideal customers, namely larger companies that are always running digital advertising campaigns. The strategy seems to have worked, with Instapage not losing a single enterprise customer in the past year.

Quick also sees a broader opportunity around what he calls “post-click conversion,” which he said has three key components — scalable creation, optimization and personalization. (That last category is what his team is focused on building out now.) And while other companies like Optimizely might tackle some of those segments, he said Instapage is leading the way at combining all three.

I wondered whether GDPR or other shifts in the landscape around online data and privacy could affect Instapage’s business, but Quick argued that the company doesn’t need detailed user data to be helpful: “The personalization comes form the advertising itself. We can glean insights from what ads did they engage with.”

As for raising money, Instapage had already raised $2 million in seed funding. Quick said that when he started to have conversations about a Series A, most VCs weren’t interested in giving the company “the valuation we felt we deserved.” (He speculated that they were “a little preoccupied by the buzzy things that weren’t going on” like cryptocurrency and on-demand delivery.) And the company was profitable, with what Quick said is $12 million in annual recurring revenue, so it didn’t need to raise the funding.

However, he found that Morgan Stanley was more willing to dig in and learn about Instapage’s business.

“Instapage has a clear vision for improving digital marketing efficiency and has built an impressive product for marketers to create more successful advertising campaigns without the need for external or technical resources,” said Pete Chung, the head of Morgan Stanley Expansion Capital, in the funding announcement. “We’re confident that this next phase of growth will further establish Instapage as the authority on converting ad clicks and optimizing conversions, and we look forward to working with Instapage to bring post-click optimization to a larger audience.”

Walmart is expanding its grocery delivering business with the help from new partner, Postmates, the retailer announced this morning. Postmates will initially begin Walmart Online Grocery Delivery in Charlotte, North Carolina with plans to reach other markets in the months ahead. Postmates joins Walmart’s existing delivery partners Uber and Deliv, who have been helping Walmart test deliveries in select markets, including Dallas, Denver, Orlando, Phoenix, Tampa and San Jose.

The addition of a new delivery partner is not unexpected – the company recently detailed its plans to expand its grocery delivery business across the U.S. in 2018, going from just 6 metros to 100 during that time. That will see Walmart offering delivery in 800 of its stores, and able to service over 40 percent of U.S. households.

To date, Walmart has been working with partners instead of rolling out its own delivery service  because it wants to be able to quickly scale up its grocery delivery operation, given the looming threat of Amazon and its own rapid expansion of Whole Foods grocery deliveries through Prime Now.

In fact, today Amazon announced Whole Foods grocery delivery has reached yet another major metro – Los Angeles and Orange County. Amazon has been moving fast to integrate Whole Foods into its Prime business, with everything from deeper discounts and coupons for Prime members to the option to have Whole Foods purchases delivered. And it has taken advantage of Whole Foods stores for its Amazon Lockers, selling its own electronics, and more.

Walmart hadn’t initially been focused on grocery delivery, touting that pickup was a more affordable option for its customer base due to the high costs of delivery. But it couldn’t let Amazon dominate the grocery delivery market either. However, Walmart’s Curbside Pickup is much more widely available and will be for some time. The service is now live at 1,200 Walmart stores, and expected to roll out to 1,000 more in 2018.

Walmart’s delivery service does offer a competitive advantage over some rivals, including Instacar and Shipt – it doesn’t mark up the cost of goods, nor does it require (or offer) a subscription. Instead, there’s only a $9.99 per-order fee, which helps to fund the partner-provided delivery.

The addition of Postmates as a Walmart partner could be beneficial to the delivery provider, too, which has been looking into ways it could better compete with the likes of Uber and Amazon. The company has also been struggling to reach profitability, posting an operating loss of $75 million for 2017 on revenues of $250 million, Recode recently reported. It has also discussed a sale with GrubHub, said the report.

“Both Walmart and Postmates strive to make the lives of our customers easier,” said Dan Mosher, senior vice president, merchant lead at Postmates, in a statement about the Walmart partnership. “With our growing fleet of 160,000 couriers, we are confident that we’ll be providing Walmart customers with the ultimate convenience.”

Walmart didn’t say when Postmates would expand beyond Charlotte, only that it would “soon.”

The pay gap is real, my friends. Hired, the career marketplace that produces studies about pay inequality, spent $2 million in 2016 to close its wage gap across gender, race and sexual identity. Since then, Hired says it reassessed pay six months later, and followed the same methodology in 2017. The company says it plans to do it again this year.

It’s worth noting that Salesforce, which employs 25,000 employees or so, spent $3 million in one year to close its gender pay gap. In total, Salesforce has spent about $6 million to close the gender and racial pay gap since 2015. Hired, in contrast, employed only 200 people at the time, and had to give raises to 60 percent of its employees. Half of those employees were women.

“This is because we didn’t have a hiring and compensation model in place early on, not unlike many fast growing start ups,” Hired SVP of People Kelli Dragovich wrote in a blog post. “We tended to pay people based on what they asked for in the hiring process and what their previous salary was, instead of their objective market value. Now, we base our compensation decisions on third-party salary data that takes skillset, years of experience, geographic location, industry and company size into account.”

Moving forward, Dragovich says Hired intends to continue cross-checking wages with market salary data and analyzing employee salaries. Today, Hired pays women more money than men in three of the company’s four core functions of engineering, marketing, talent advocacy and sales.

“By sharing our own journey, we hope that more companies will understand that they too can be part of the solution,” Dragovich wrote.