Monthly:December 2017

2017: DeepMind’s year in review

In July, the world number one Go player Ke Jie spoke after a streak of 20 wins. It was two months after he had played AlphaGo at the Future of Go Summit in Wuzhen, China.After my match against AlphaGo, I fundamentally reconsidered the game, and now I can see that this reflection has helped me greatly, he said. I hope all Go players can contemplate AlphaGos understanding of the game and style of thinking, all of which is deeply meaningful. Although I lost, I discovered that the possibilities of Go are immense and that the game has continued to progress.Read More

Collaborating with patients for better outcomes

Working as a doctor in the NHS for over 10 years, I felt that I had developed good understanding of how patients and their families felt when faced with an upsetting diagnosis or important health decision. I had been lucky with my own health, having only spent one night in hospital for what ended up being a false alarm. But when my son was born prematurely two years ago, I had a glimpse into what being on the other side feels like – an experience that has profoundly shaped my thinking today.It wasnt until I was waiting to hear, rather than give, important health updates that I really understood the feeling of uncertainty and powerlessness that many patients and their families feel. It really put into perspective how important it is to involve patients, and their families and carers, in their own health – that care is not something done to a patient, but rather, something that is shaped by everyone involved in the healthcare process.In my first week at DeepMind Health, I was really impressed that one of my new colleagues (not a nurse or doctor) had set up a meeting so we could hear directly from a patient, Michael Wise, who ended up needing dialysis and a kidney transplant after a sudden and unexpected problem with his kidneys. Since then, weve continued to increase our efforts to bring the patients voice into our projects.Read More

Discover the true value of each click with data-driven attribution

From buying a TV to planning a holiday vacation, it takes consumers, on average, four clicks before completing a transaction online.¹ Some of these clicks matter more than the others, yet 40% of businesses still use first-touch or last-touch attribution when measuring ad performance.²
In order to understand the true value of each of those clicks, you need an attribution model that better reflects the new fragmented mobile-first consumer journey. Data-driven attribution (DDA) in DoubleClick Search solves this by helping you better understand the value of each ad click throughout your customer’s journey.

Create a model that fits your needs

Machine learning sets DDA in DoubleClick Search apart from rules-based attribution models. Unlike those models, DDA calculates the value of each ad click using your Floodlight conversion data, including conversions that happen across devices and browsers. By looking at all the clicks that happen across all those conversions, DDA can assign more value to the clicks that are driving conversions, and less value to those that don’t.
Over time, your model will become more accurate as more data becomes available to inform your bids with. You can also customize your DDA model to fit your business goals.

DDA gets results

Here’s how DDA has already created value for advertisers around the world:

“Data-driven attribution has taken our campaign performance to the next level. By creating tailored attribution models for both new and existing customer journeys, we’re able to easily see which campaigns are contributing to our chosen business objective,” says Matt Darbon, Online Marketing at Waitrose. The British supermarket chain decreased its cost-per-order by nearly 40% while significantly increasing sales volume.

Walks of Italy provides immersive tour experiences around Italy’s major cities and attractions. Using DDA and Smart Bidding, the brand saw a 31% increase in ticket sales, compared to 11% for other non-brand campaigns not running on DDA bid strategies. It also saw 25% growth in overall return on investment (ROI).

Volkswagen Group, one of the world’s leading vehicle manufacturers, saw sales leads increase by nearly 6% and cost-per-lead decrease by 17% in Brazil with DDA.

TalkTalk, the UK’s leading value-for-money telecoms provider, partnered with mSix to implement DDA for its campaigns. As a result, the brand saw its total volume of sales increase by 9% while maintaining its cost-per-acquisition.

Make DDA work for your business

Here are some best practices to help ensure your data-driven attribution models are customized to meet your business goals:

  • Apply your customer and market knowledge. Create and apply DDA specific labels to your campaigns based on the role that Search activity plays in the customer journey. Learn more
  • Check the accuracy of your model with Floodlight columns. Once you’ve created your custom DDA model, create Floodlight columns for DDA conversions to ensure your model is set up correctly. Learn more
  • Track your performance against other models. For easier monitoring, add your DDA model to your favorite reports. Create a custom column that shows the percentage difference in conversion value assigned to keywords, campaigns and labels between your attribution models. Learn more
  • Re-evaluate keywords, campaigns and labels for performance. Now that you’re measuring the value of every click, you may discover that some of your campaigns are performing better than before. Make sure your budget and bids are updated based on the insights provided by DDA. Learn more
  • Automate your bids based on your new model. To get the full benefits of DDA, use Smart Bidding with your DDA model to automatically set optimal bids across your campaigns. Learn more

Learn more about data-driven attribution in the DoubleClick Search Help Center.

¹ Google Analytics MCF data for single device measurement, Feb. 2017
² Google/Forrester, “Cross Channel Attribution is Needed to Drive Marketing Effectiveness,” May 2014, U.S.

Posted by Michel van Luijtelaar
Measurement & Attribution Specialist, DoubleClick SearchRead More

Hangouts Meet now available in the Google Calendar API

Thousands of developers use Google Calendar API to read, create and modify Google Calendar events, and quite often, these events represent meetings happening not just face-to-face but also remotely. We introduced Hangouts Meet earlier this year to give users richer conference experiences, adding video call links and phone numbers for G Suite Enterprise. Starting today, we are making it possible to access all that conference information through the Google Calendar API. With this update, developers can now:

  • Read conference data associated with events 
  • Copy conference data from one event to another 
  • Request new conference generation for an event 

The API supports all Hangouts versions.

Reading conference data 

Conference information is stored in a new event attribute called conferenceData.conferenceData provides information about the solution that was used to create the conference (such as Hangouts Meet) and a set of entry points (like a video call link and phone number). Everything the user needs to know to join a conference call is there.

To help you build even better user experiences, we also give you access to icons and user-readable labels that you can use in your products. In JSON format, conferenceData looks something like this (of course, your actual meeting IDs and phone numbers will vary):

And this, for example, is how you would retrieve and display conference solution name and icon:

The result of the code above will look like this in the user interface:


Copying conferences across events 

Sometimes displaying information is not enough—you might want to update it as well. This is especially true when scheduling multiple Calendar events with the same conference details. Say you’re developing a recruiting application that sets up separate events for the candidate and the interviewer; you want to protect the interviewer’s identity, but you also want to make sure all participants join the same conference call. To do this, you can now copy conference information from one event to another by simply writing to conferenceData.

To ensure that only existing Hangouts conferences are copied, and to help safeguard your users against malicious actors, copied conference data will always be verified by the Google Calendar API using the signature field, so don’t forget to copy it too.

Creating a new conference for an event 

Finally, the API allows developers to request conference creation. Simply provide a conferenceData.createRequest and set the conferenceDataVersion request parameter to 1 when creating or updating events. Conferences are created asynchronously, but you can always check the status of your request to let your users know what’s happening. For example, to request conference generation for an existing event (again, your request and event IDs will be different):

The immediate response to this call might not yet contain the fully-populated conferenceData which is indicated by status pending:

Once the statusCode changes to success, the conference information is populated. Finally, if you are developing a Google Calendar client, you might also want to know beforehand which of the three Hangouts solutions (consumer Hangouts, classic Hangouts and Hangouts Meet) will be used to create the conference. You can get that information by checking allowedConferenceSolutionTypes in a calendar’s conferenceProperties.

To get started, check out the documentation page for managing conference data. We can’t wait to see what you build with these new features in the Google Calendar API.

Read More

A look at Team Drives in action at the California Academy of Sciences

Located in San Francisco’s Golden Gate Park, the California Academy of Sciences is an aquarium, planetarium, and natural history museum all wrapped into one. Attracting visitors from all over the world, the California Academy of Sciences aims to explore, explain, and sustain life on Earth. In addition to biodiversity research and conservation efforts, they offer a large variety of exhibits to educate visitors about wildlife, ecosystems, and the sustainability of our planet.

The California Academy of Sciences uses G Suite and other Google products to help employees collaborate, onboard new team members effectively, manage data for science-based animal care, and schedule upcoming physicals and treatments for live animals. Recently, they migrated all their digital data to Team Drives, a G Suite for Nonprofits tool that lets organizations store, search, and access shared content from anywhere. In Team Drives, files belong to the team instead of the individual, so users won’t need to search across siloed folders with varying permissions. Since implementing this change, the California Academy of Sciences has been able to reduce time spent searching for documents, limit duplication of efforts, and collaborate more closely with their team members and other organizations internationally. We spoke with Associate Director of the Steinhart Aquarium (and Google super user), Laurie Patel, who successfully migrated 15 years of digital data to Team Drives in just one evening, to learn more about how they’re using the tool.

Three endangered African Penguins on exhibit at the California Academy of Sciences. The penguin on the left is a juvenile hatched at the Academy as a part of the Association of Zoo and Aquarium Species Survival Plan.
Three endangered African Penguins on exhibit at the California Academy of Sciences. The penguin on the left is a juvenile hatched at the Academy as a part of the Association of Zoos and Aquariums’ (AZA) Species Survival Plan.

Better animal health management through unlimited storage

The initial reasoning behind the transition to Team Drives was unlimited storage. Because of the massive amounts of animal medical data that must be stored, the aquarium team needs space to upload all the PDFs, images, videos, and spreadsheets that they collect. All medical data gets logged, like each animal’s annual physicals, blood work, pictures, weight, and other diagnostics. With 38,000 live animals at the California Academy of Sciences, it’s easy to see how the virtual file cabinet of data in their systems could start to overflow. With Team Drives, Laurie’s team can upload all the images and data they collect so that it’s accessible in one place, all the time—without relying on an individual owner to have sole access. And with Team Drives’ permissions settings, they share and link these folders to the external Zoological Information Management System (ZIMS) database. That database connects with zoos and aquariums across the world so researchers can cross-reference each species’ baseline health reports. Being able to upload large files to this database has increased both the California Academy of Sciences’ and the ZIMS accumulated knowledge of medical data to ensure all animals are treated properly and receive the best possible care and enrichment.

Everything you need to know is right there in Team Drives. Laurie Patel
Associate Director of Steinhart Aquarium

Real-time updates to support strict protocols for animal safety

Caring for a diverse animal collection in varied habitats, like the four-story Osher Rainforest exhibit or the 212,000 gallon Philippine Coral Reef exhibit, requires California Academy of Sciences’ staff to adhere to strict protocols to ensure a consistently high standard of animal care. To ensure stable environments, all processes need to be executed in a specific way—and this critical information has to be readily accessible to staff and always up to date. From changing an animal’s diet to venomous animal handling protocols, employees routinely search and access these procedures and databases to make real-time decisions. For example, water is collected daily from separate tanks to check the water quality and test things like pH levels and magnesium concentration. Employees input this data into Google Sheets, and conditional formatting automatically attributes a color code based on each test result—an easy and instantaneous visual indication to inform what action is needed for the employees back at the tank.

We use data-driven responses for science based animal care. And utilizing Google’s collaboration tools for all this data is how we’re able to do this. Laurie Patel
Associate Director of the Steinhart Aquarium

Streamlined onboarding = more time for animals

By consolidating all training materials and important resources in one place, the Steinhart Aquarium team can onboard new members to the team quickly and efficiently. This helps the team prepare for legacy planning as well. When one teammate leaves, their successor can easily take ownership of all the files and resume where the former employee left off, ensuring that no work is lost in the transfer. And by linking to various Team Drives folders in their online hub powered by Google Sites, they’ve created a one-stop-shop to guide team members to the right information at the right time.

The Reef Lagoon exhibit showcases the interrelationships of mangrove, lagoon and reef habitats found in the Philippines, an ecosystem researched by the biologists and scientists at the California Academy of Sciences.

Ultimately, Team Drives help California Academy of Sciences operate without fear of lost data or out-of-date sharing preferences. This extra time saved allows employees to spend more time caring for a charismatic group of live animals and engaging with the museum’s visitors, rather than their screens. Learn how Team Drives can help your organization and get started today.

Read More

Google Drive recognized as a leader by Forrester for File Sync and Share (EFSS) cloud solutions


We’re designing cloud-based solutions to help teams become more knowledgeable and more effective. Businesses like Whirlpool are turning to Google Drive because they need more secure, friction-free tools to help them house—and more importantly, share—collective company knowledge.

Our continued efforts to help businesses collaborate in the cloud have not gone unnoticed. Today, Drive has been recognized as a leader in The Forrester Wave™: Enterprise File Sync And Share Platforms–Cloud Solutions, Q4 2017 report, which ranks cloud-based file sharing and storage solutions for enterprise IT professionals. Earlier this year, Drive was also named a leader among content collaboration platforms by Gartner.

What does this mean for your business?

There are inherent benefits to working in the cloud, and businesses are starting to take advantage of them. Cloud-first tools like Drive gives companies a single place to store, access and share information, which can improve efficiencies.

Of course, you need to be confident that these new cloud solutions will work well with your existing IT investments. Over the past year, we’ve enhanced Drive with additional enterprise features to make migrating from on-prem fileshares and content management systems easier:

  • Team Drives, so you can work on collectively-owned files and feel confident that your work will stay put no matter who comes or goes on your team. 
  • AppBridge, so your enterprise can migrate all of its data more securely to the cloud with ease. 
  • Drive File Stream, to help you access all data quickly without using up a ton of disk space.  
  • Data Loss Prevention and Google Vault for Drive, to help you control how your teams share information and to help you keep sensitive data safe.
  • Quick Access in Drive, to help surface your data faster with artificial intelligence-powered technology.

To learn more, access the 2017 Forrester Wave EFSS Cloud Solutions report here. Or get in touch with us to learn more about how your business can benefit from moving to Google Cloud.

Read More

Pivot to the cloud: intelligent features in Google Sheets help businesses uncover insights

When it comes to data in spreadsheets, deciphering meaningful insights can be a challenge whether you’re a spreadsheet guru or data analytics pro. But thanks to advances in the cloud and artificial intelligence, you can instantly uncover insights and empower everyone in your organization—not just those with technical or analytics backgrounds—to make more informed decisions.

We launched “Explore” in Sheets to help you decipher your data easily using the power of machine intelligence, and since then we’ve added even more ways for you to intelligently visualize and share your company data. Today, we’re announcing additional features to Google Sheets to help businesses make better use of their data, from pivot tables and formula suggestions powered by machine intelligence, to even more flexible ways to help you analyze your data.

Easier pivot tables, faster insights

Many teams rely on pivot tables to summarize massive data sets and find useful patterns, but creating them manually can be tricky. Now, if you have data organized in a spreadsheet, Sheets can intelligently suggest a pivot table for you.

In the Explore panel, you can also ask questions of your data using everyday language (via natural language processing) and have the answer returned as a pivot table. For example, type “what is the sum of revenue by salesperson?” or “how much revenue does each product category generate?” and Sheets can help you find the right pivot table analysis.

Pivot GIF #2

In addition, if you want to create a pivot table from scratch, Sheets can suggest a number of relevant tables in the pivot table editor to help you summarize your data faster.

Suggested formulas, quicker answers

We often use basic spreadsheet formulas like =SUM or =AVERAGE for data analysis, but it takes time to make sure all inputs are written correctly. Soon, you may notice suggestions pop up when you type “=” in a cell. Using machine intelligence, Sheets provides full formula suggestions to you based on contextual clues from your spreadsheet data. We designed this to help teams save time and get answers more intuitively.

Formula suggestions in Sheets

Even more Sheets features

We’re also adding more features to make Sheets even better for data analysis:

  • Check out a refreshed UI for pivot tables in Sheets, and new, customizable headings for rows and columns.
  • View your data differently with new pivot table features. When you create a pivot table, you can “show values as a % of totals” to see summarized values as a fraction of grand totals. Once you have a table, you can right-click on a cell to “view details” or even combine pivot table groups to aggregate data the way you need it. We’re also adding new format options, like repeated row labels, to give you more fine-tuned control of how to present your summarized data.
  • Create and edit waterfall charts. Waterfall charts are good for visualizing sequential changes in data, like if you want to see the incremental breakdown of last year’s revenue month-by-month. Select Insert > Chart > Chart type picker and then choose “waterfall.”
  • Quickly import or paste fixed-width formatted data files. Sheets will automatically split up the data into columns for you without needing a delimiter, like commas, between data.

These new Sheets features will roll out in the coming weeks—see specific details here. To learn more about how G Suite can help your business uncover valuable insights and speed up efficiencies, visit the G Suite website. Or check out these tips to help you get started with Sheets.

Read More

DeepMind papers at NIPS 2017

Between 04-09 December, thousands of researchers and experts will gather for the Thirty-first Annual Conference on Neural Information Processing Systems (NIPS) in Long Beach, California.Here you will find an overview of the papers DeepMind researchers will present.Read More