Trust. And safety

Yelp changed its algorithm and launched a site to increase transparency. Speaking of transparency...


It’s March, and this month I’m running a subscription drive of sorts in order to grow my paid audience. Why should you pay for a newsletter I send for free? Here’s my brief sales pitch. Also: Beginning in April, paid subscribers will receive this newsletter 24 hours before free subscribers. Paid issues will arrive on Wednesdays, as usual. Independent restaurant owners and employees are eligible for free “paid” subscriptions. If you can (and if you haven’t already!):

Upgrade your subscription


Last week, Yelp released its 2020 trust and safety report and launched a site to support it. Turns out, about a quarter of the 18.1 million reviews Yelp received last year were flagged for suspicious activity by Yelp’s recommendation software. The software itself has been developed over the past 15 years, according to the company, to showcase the reviews it deems most helpful. Also last week, Yelp rolled out an update to its recommendation software that it says will cause approximately 177,000 currently recommended reviews at 154,00 businesses will become not recommended. Yelp has over 224 million contributed reviews to date.

Public complaints from businesses about Yelp’s algorithm usually focus around reviews deemed somehow too positive to post. Yelp says the un-recommended reviews generally represent posts that are “a conflict of interest, fake, less useful, solicited, or otherwise less reliable.” Often, these reviews come from what Yelp calls “review exchange groups” or “review rings” that offer to sell or trade fraudulent reviews. In 2020, Yelp says it closed over 1,200 user accounts for being associated with these review rings, and the software update  is aimed directly at reducing these instances of fraud. 

The other big standout for me:  Yelp said instances of what it calls media-fueled incidents nearly tripled in 2020. That is: people leaving reviews of a business because it was in the news, not because they had a legitimate experience at that business. There’s a host of other stats from last year on Yelp’s new trust and safety page. It’s nice that Yelp is providing the receipts here to increase transparency — even just a little bit — into its practices. It’s good that the company is shutting down the bad actors, and even better that it’s going after more of them than ever. 

But let’s talk about trust in restaurant tech. Last week, a high-profile Los Angeles Times story made a lot of people angry. The story was that a fraudster duped a restaurant out of over $700 of food. The response was a mix of “how is this even possible?” and “people suck,” though sadly the restaurant response more closely resembled, “Yep, it’s true.” In fact, I’ve heard via a number of sources that although fraud has always been a problem in the business (the LA Times headline referred to these scams as the new dine-and-dash), they’ve seen an uptick in fraudulent behavior since the pandemic began. The story that started it all involved Tock because that’s where the fraudulent $728 order was placed; the dispute involved the rightful credit cardholder and their bank — not Tock. (I covered this in greater detail last week, ICYMI.) 

In the week since, more stories have emerged about the challenges some businesses face working with restaurant technology companies including this one from Nom Wah’s Barb Leung about a $641 Grubhub liability for unconfirmed orders and the resulting back-and-forth between the businesses. 

This reminds me a bit of pre-Covid restaurant technology days when the big debates centered around the ability of a third party like Grubhub to accurately represent, and in a sense, advocate for a restaurant on its platform. (To be clear, my evidence is anecdotal in the form of stories repeated to me and I haven’t pursued the issue with Grubhub.) Now, a year into a pandemic, the issue here has shifted a bit: not only does a restaurant need to worry whether or not its choice of tech vendor is representing its own best interests, it needs to trust that the tech provider is keeping it safe. It’s probably time for more of them to show the receipts. 


McDonald’s might offload parts of Dynamic Yield

McDonald’s is considering selling part of Dynamic Yield, the AI startup it purchased for over $300 million in 2019. According to the Wall Street Journal, McDonald’s store owners have complained that the predictive technology hasn’t delivered on its promised sales results. In a Friday statement, Dynamic Yield’s founder said that a partial sale was always a possibility, and “now feels like the right time to explore that possibility.” 

McDonald’s is said to be looking into selling the parts of Dynamic Yield that work with other businesses, which would, clearly, make sense. 

I remember when this acquisition happened (and not just because I got reprimanded for not adding the sale price to the headline, though I’ll never forget the $300 million-plus price tag, I promise!) — it felt pivotal because it gave startups and other small restaurant tech companies another place to land. I started asking the leaders of any startup I talked to if they’d sell to McDonald’s, and the answer was usually off the record and some version of, “Yeah, of course we would.” 

In the years since this acquisition, large restaurant brands who have grown dependent on technology for growth and customer satisfaction, have continued to develop in-house technology. Acquisitions of this sort are certainly not off the table, though acquisitions of this size could be, if this story is any indication of how they’re going. It also raises important questions about the future: will tech companies build for a brand or for the business? What will that mean?

It’s an idea I’d like to explore further, especially since several tech companies have recently staked their business on some version of “we’re democratizing [big company X]’s technology for small businesses.” My biggest question for those companies is: What’s your exit strategy? 


What else is happening?

DoorDash held its first quarterly earnings call last week, and the good news is that its revenue more than tripled in the fourth quarter, stretching to $970 million. The bad news: the company lost $312 million overall. 

“Third-party delivery is not profitable,” has been the headline of choice for nearly a year as restaurant operators contend with challenge after challenge. Last week, Grubstreet ran a piece titled Big delivery is winning — even if everyone hates it. The week before, the same publication ran the ghost kitchens are dumb piece. 

In a recent Twitter thread, I laid out why I find this line of thinking short-sighted. Granted, the average consumer isn’t thinking about what’s going to happen in restaurant technology six months down the road; they’re hoping their favorite restaurants make it through the next four to six weeks as we all wait for the dim light at the end of the tunnel to get brighter. But in its earnings call, DoorDash said that new costs from temporary delivery fee caps and California’s Proposition 22 cost the company $36 million in revenue in the fourth quarter — an amount large enough to push the company past the $1 billion mark had these hurdles not existed. 

DoorDash has made a number of investments in its own post-pandemic future, including moves in ghost kitchens and, of course, Sally the salad robot. By its own admission, the company’s goal is to become the first stop for a restaurant that wants to grow its business. Some analysts and journalists wonder if DoorDash can persist through a successful vaccine rollout and a return to normal life — this journalist is wondering how much it’ll continue to grow on the other side. 


Instacart raised another $265 million at a $39 billion valuation. As TechCrunch notes, the valuation is over double its previous: $17.7 billion when it raised $200 million in October 2020. That’s also $9 billion higher than a rumored IPO valuation last November. The Information’s coverage notes a potential direct listing might instead be in the cards for Instacart — it added a board member formerly of Spotify, one of the higher profile direct listings of recent memory. But, The Information also has thoughts on timing: “The longer Instacart waits, the more we’ll get closer to a post-pandemic era when grocery deliveries become more of a luxury than a necessity.” 


Uber officially spun out its robot startup. What was Postmates X is now Serve Robotics and is reportedly working to close a series A round of funding, per Techcrunch.


Wix acquired SpeedETab. The website provider acquired an online ordering company to create a more comprehensive product for restaurant customers, according to an SVP at Wix. Wix says that the number of restaurants using its platform more than doubled over the last two years, and the number online orders nearly quintupled. 



Delivery company Hungr launched some virtual brands, all based on data. The concepts include a pizza brand, a burger brand, and a barbecue brand and used customer search data to help shape the menus. They’re available exclusively on the Hungr platform.