All posts in “Advertising Tech”

Facebook has a 100-person engineering team that helps advertisers build tools and infrastructure

Stories about Facebook’s advertising business tend to focus on the big numbers — its billions of users, millions of advertisers or its enormous lead over any competitor that’s not named Google.

But Facebook says that one of its success stories in recent years involves a relatively small group of engineers — in fact, originally it was just one engineer, Vastal Mehta, who serves as Facebook’s director of solutions engineering and now leads a team of more than 100 people. That team works with advertisers to build the technology and infrastructure needed to run more effective campaigns on Facebook, often on top of Facebook’s APIs.

Mehta said that when he first started working on this in 2010, it was a very different landscape, both for mobile (where BlackBerry was still a major player) and for Facebook (which hadn’t even introduced advertising into the News Feed). This was right as the company was trying to shift in a big way toward mobile, and advertisers were still trying to wrap their heads around the change: “For example, travel companies didn’t have teams set up to reach consumers with mobile advertising. We knew that we needed to invest in helping businesses build infrastructure to power their mobile advertising, so I started a team that could help businesses in this sort of bespoke way,” Mehta added.

You’d expect any digital media business to offer some degree of technical support to its biggest advertisers, but the solutions engineering team is actually building products.

For example, it was involved in creating Facebook’s dynamic ads format (where ads show different products to different users based on their activities and interests). Mehta said dynamic ads were first inspired by the complaints of an advertiser he was meeting with in Hamburg, Germany, and he then worked with the Facebook Ads team to create a prototype, eventually leading to a more polished product and broader availability.

It’s probably safe to say that not every client meeting leads to a new ad format — sometimes Mehta’s team is just helping advertisers understand how to use their existing tools in a more effective way. But that other option, working with the rest of Facebook to build something new, is also on the table.

Vatsal Mehta

To give me a better sense of what the team actually does, Facebook connected me with Anthony Marino, chief marketing officer at online thrift store thredUP. Marino said that when his company started talking with Facebook’s solution engineering team in 2016, there was a big challenge: How to use ads to highlight thredUP’s constantly changing inventory.

“On thredUP, the site is practically remade every hour as thousands and thousands of new items are added,” he said. “We looked at that flow of product, of apparel, and it was like being a news site … We had to figure out a way to automate the process of, okay, once we capture and the attributes and qualities of different items of clothing, how do we get them in front of the right person?”

To enable that, Facebook worked with thredUP to launch dynamic ads that were connected to thredUP’s real-time product catalog. The system uses machine learning to further improve the targeting; for example, showing users different types of ads at different times of day.

“The first thing is, Facebook puts the right people in the room,” Marino said. He recounted working with Facebook to create “new ad products, new data pipelines” between the two systems, and he said, “There were product people, there were operations people in the room. We were able to really integrate at the data integration, at the business process level.”

Did this actually lead thredUP to buy more ads on Facebook? The companies didn’t share numbers about the company’s ad spending, but part of the process involved shifting thredUP from Criteo retargeting to Facebook dynamic ads, and Marino told me, “Working with the solutions engineering team at Facebook enabled us to spend our dollars more efficiently, so that we could amp our marketing budget and drive more new customers to”

The team has worked with other customers, including Michael Kors, Edmunds, The New York Times, Gilt and Zynga. It also works with the companies that offer ad-buying tools on top of Facebook, like Smartly, Kenshoo, Marin Software, Adobe, Social Code and Nanigans — and mobile gaming company Machine Zone said the Facebook ad-buying platform it built with the solution engineering team’s help was so successful that it’s launching a new business called Cognant.

Facebook says that on average, clients working with the solutions engineering team see their return on ad spend improve by 100 percent.


Of course, while Facebook continues to do extraordinarily well financially, it’s been battered in public perception as the government scrutinizes the role it may have played in spreading misinformation as part of Russia’s election interference efforts. On the ad side, Facebook has announced new transparency features like the ability to see every ad campaign from a given advertiser, and an archive of ads related to federal elections.

When I brought this up, the company said these changes, and the broader political environment, haven’t really affected the day-to-day work of the solutions engineering team, which is much more in the trenches, helping advertisers do new things.

As for what they’ll be up to in 2018, Mehta said:

One area we’re increasingly spending time helping clients with is incorporating more machine learning into solutions and driving efficiency through technology. This includes building better optimization tools that help the client without them needing to adjust and turn nobs in the interface. We see this as a huge area of investment across our business over the next year.

Featured Image: Facebook

TrendKite raises another $11M for its PR analytics platform

TrendKite is closing out 2017 with the announcement that it’s raised an additional $11 million in funding.

The Austin-based startup says it’s currently analyzing 4.2 million articles every day, and using that analysis to help both brands and agencies to measure the impact of their PR efforts. So it can identify the total audience reached by articles mentioning the company, or highlight which articles had the most impact on brand awareness.

TrendKite’s revenue has grown by more than 100 percent year-over-year, with customers including Mondeléz International, Nike and Delta. The company says the new funding will go towards additional product development, with the goal of making “PR software as indispensable to the CMO as CRM or marketing automation software.”

Touting the value of “earned media” (namely, getting attention via editorial content rather than advertising) might seem a little less convincing at a time when the President of the United States has repeatedly attacked mainstream media outlets, and there’s more concern that ever that everyone’s getting stuck inside their own filter bubbles.

Asked via email how TrendKite’s role will have to evolve in the current environment, CEO Erik Huddleston argued that the these broader changes have mostly been limited to politics: “A thoughtful review on the latest smart phone by a trusted thought leader is still much more powerful than a banner ad for the same phone or the marketing copy on the product page.”

And if that changes?

“If it does, my hunch is that the value of TrendKite and platforms like it only increases as identifying the right publications and journalists that have influence over your specific target market will become harder with increased ‘partisan’ media affiliation and understanding the resulting impact will likewise take on increased complexity,” Huddleston said.

This round brings TrendKite’s total funding to more than $46 million. It was led by Harmony Partners, a firm whose portfolio includes Chartbeat, mParticle and Postmates.

“TrendKite’s high growth and operational excellence, combined with the market opportunity in making earned media relevant to the C-suite, make this an obvious choice for Harmony Venture Partners,” said founder and managing partner Mark Lotke in a statement.

Featured Image: cifotart/Getty Images

Tech in 2017: Crazy, troubled and out of control?

There’s no doubt about it, it’s been a difficult year for the technology industry.

It’s also been a crazy year for tech, with cryptocurrencies surging and rollercoastering in value — presumably minting a few millionaires along the way, assuming they actually cashed their coin out.

Yet even those strange crypto-highs have come with some equal and opposite lows: Scams, hacks, pointlessly wasteful energy consumption, and — zooming out — the baffled confusion of anyone outside the crypto-boom trying to understand what logic (if any) it runs on.

Meanwhile AI’s touted efficiency gains have also been weighed down by counternarratives this year — the stories where AI is shown preferring to promote clickbait. Or to be biased. Or worse.

AI blowing “bubbles of hate” as one YouTube policy staffer memorably put it during a political grilling. Though he merely advocated the use of yet more AI to fix the problem.

The naivety and irresponsibility of tech giants whose platforms have scaled so big and got so powerful left the strongest impression in 2017, as marketing claims about fostering ‘openness and connection’ unravelled in the face of the literal opposite: Rising division and social strife.

If only Twitter had listened to all the users telling it to fix its troll problem for years. Ditto YouTube and its below-the-fold comment hellscape. And Facebook and fake news.

Worse was on show too: Uber’s reputation lies in tatters for a reason. And massive data breaches came to seem like an almost routine occurrence in 2017.

Sexism and sexual harassment were also shown to be an ugly and embedded problem across the industry.

Going into 2018, tech certainly has a lot of cleaning house to do.

Looking ahead, companies of all sizes should be trying to see outside their own filter bubbles and Kool-Aid-stocked canteens — and asking themselves genuinely tough questions about who and what will be impacted by their technology.

The unpopularity being sharply directed at once shiny tech brands should be a paradigm shifting wake up call.

Questions are being asked about platform power. Regulatory rules and knives are being sharpened. Politicians are eager to point the finger of blame. And with so much tech-fueled ammunition, who can blame them?

After the surge, the crash.

Tap the arrows to take an A to Z tour through the tech that troubled the news this year.

EyeEm’s new products aim to understand brand aesthetics

EyeEm is unveiling new tools to help the brands and marketers using the site to source their images.

Underlying these tools is a technology called EyeEm Vision, which we described in-depth earlier this year. The goal is to expand image recognition so that it’s not just identifying the objects in the photo, but also its aesthetic qualities.

EyeEm’s co-founder and chief product officer Lorenz Aschoff described EyeEm Vision as an extension of the photography marketplace’s broader mission to address “the content crisis” — namely the fact that when EyeEm was founded in 2011, Aschoff felt that there was a “massive flood of images” that had “completely destroyed the visual aesthetics of the web.”

EyeEm aims to fix that by helping brands find beautiful photographs. And Aschoff said EyeEm Vision has been trained to identify many of the visual elements that make for a good photograph — it is, in his words, “technology that understands, in general, beauty.”

At the same time, he acknowledged, “What I think is beautiful might be different from what you think is beautiful.” Plus, individual brands are going to have their own specific standards and guidelines that go beyond beauty. So each customer can upload photos that train EyeEm Vision to identify photos that match their own aesthetic — Aschoff said EyeEm’s analysis is looking at around half a million different factors.

EyeEm personalized search

One of the ways EyeEm is actually deploying the technology is by launching a new Missions Dashboard. Brands use Missions to crowdsource campaign photos from the EyeEm community, and the new dashboard allows them to track how their Mission is going — how many photographers are participating, how many photos have been uploaded and so on. EyeEm says that the average Mission results in more than 100,000 photos, which why it’s important to use EyeEm Vision to surface the photos that best match the brand’s style.

EyeEm is also incorporating Vision into a personalized search product, where marketers can search the EyeEm image library, filtered based on their own brand guidelines. For example, BCG’s 11,000 consultants can now search for images to use in their presentations and marketing materials, and EyeEm will only show the images that are a good fit with the BCG brand.

And while this is less directly related to Vision, EyeEm is also announcing a new program called Custom, where brands can work with EyeEm photographers on custom shoots.

Lastly, if you’re curious about Vision, you can try it out for yourself on the EyeEm website.

Featured Image: Eunice Eunny/EyeEm

Twitter says Russians spent ~$1k on six Brexit-related ads

Twitter has disclosed that Russian-backed accounts spent $1,031.99 to buy six Brexit-related ads on its platform during last year’s European Union referendum vote.

The ads in question were purchased during the regulated period for political campaigning in the June 2016 EU Referendum — specifically from 15 April to 23 June 2016.

This nugget of intel into Kremlin political disinformation ops that were centered on the UK’s Brexit vote has been released as part of an ongoing internal investigation by Twitter into possible Russian Brexit meddling — initiated by a request for information from a UK parliamentary committee that’s investigating fake news.

The UK’s Electoral Commission, which oversees domestic election procedure and regulates campaign financing, has also written to social media companies asking them to investigate potential Russian Brexit meddling as part of an ongoing enquiry it’s running into whether the use of digital ads and bots on social media might have broken existing political campaigning rules.

Earlier today Facebook said it had identified three “immigration” ads bought by Russian backed accounts that ran ahead of the Brexit vote — which it says garnered 200 views.

However Facebook’s probe has so far only looked at paid content from Russian accounts. So it’s still not clear how much Brexit-related propaganda was being spread by Russian accounts on the platform given that content can also be freely shared with followers on Facebook.

In the US Kremlin agents were even revealed to have used Facebook’s Events tools to list and orchestrate real-world meet-ups. And in October, Facebook admitted as many as 126 million US Facebook users could have viewed Russian-backed content on its platform.

With Brexit, both Facebook and Twitter have yet to release this sort of ‘full reach’ analysis — so it’s still not possible to quantify the potential impact of Kremlin propaganda on the EU referendum vote.

A Twitter spokesman declined to answer additional questions we put to it, including asking for its analysis of the reach of the six ads — and whether or not it’s also investigating non-paid Russian-backed content (i.e. tweets and bots) around Brexit, not just paid ads.

An academic study last month suggested substantial activity on that front — tracking more than 150,000 Russian accounts that mentioned Brexit and some 45,000 tweets posted in the 48 hours around the vote.

Twitter’s spokesman also declined to share the Russian bought Brexit ads it has identified.

He did provide the following “key points” from Twitter’s letter to Damian Collins MP, chair of the Digital, Culture, Media and Sport Select Committee, which note an earlier decision by the company to ban ads from Russian media firms RT and Sputnik:

In response to the Commission’s request for information concerning Russian-funded campaign activity conducted during the regulated period for the June 2016 EU Referendum (15 April to 23 June 2016), Twitter reviewed referendum-related advertising on our platform during the relevant time period.

Among the accounts that we have previously identified as likely funded from Russian sources, we have thus far identified one account—@RT_com— which promoted referendum-related content during the regulated period. $1,031.99 was spent on six referendum-related ads during the regulated period.

With regard to future activity by Russian-funded accounts, on 26 October 2017, Twitter announced that it would no longer accept advertisements from RT and Sputnik and will donate the $1.9 million that RT had spent globally on advertising on Twitter to academic research into elections and civil engagement. That decision was based on a retrospective review that we initiated in the aftermath of the 2016 U.S. Presidential Elections and following the U.S. intelligence community’s conclusion that both RT and Sputnik have attempted to interfere with the election on behalf of the Russian government. Accordingly, @RT_com will not be eligible to use Twitter’s promoted products in the future.

Featured Image: Bryce Durbin/TechCrunch/Getty Images