Monthly Archives: June 2020

Watchful is a mobile product intelligence startup that surfaces unreleased features

June 4, 2020 No Comments

Meet Watchful, a Tel Aviv-based startup coming out of stealth that wants to help you learn more about what your competitors are doing when it comes to mobile app development. The company tries to identify features that are being tested before getting rolled out to everyone, giving you an advantage if you’re competing with those apps.

Mobile app development has become a complex task, especially for the biggest consumer apps, from social to e-commerce. Usually, mobile development teams work on a new feature and try it out on a small subset of users. That process is called A/B testing as you separate your customers in two buckets — bucket A or bucket B.

For instance, Twitter is trying out its own version of Stories called Fleets. The company first rolled it out in Brazil to track the reaction and get some data from its user base. If you live anywhere else in the world, you’re not going to see that feature.

There are other ways to select a group of users to try out a new feature — you could even take part in a test because you’ve been randomly picked.

“When you open the app, you’ll probably see a different version from the app I see. You’re in a different region, you have a different device,” co-founder and CEO Itay Kahana told me. He previously founded popular to-do app Any.do.

For product designers, it has become a nightmare as you can’t simply open an app and look at what your competitors are doing. At any point in time, there are as many different versions of the same app as there are multiple A/B tests going on at the same time.

Watchful lets you learn from the competition by analyzing all those different versions and annotating changes in user flows, flagging unreleased features and uncovering design changes.

It is different from other mobile intelligence startups, such as App Annie or Sensor Tower. Those services mostly let you track downloads and rankings on the App Store and Play store to uncover products that are doing well.

“We’re focused on everything that is open and visible to the users,” Kahana said.

Like other intelligence startups, Watchful needs data. App Annie acquired a VPN app called Distimo and a data usage monitoring app called Mobidia. When you activate those apps, App Annie captures data about your phone usage, such as the number of times you open an app and how much time you spend in those apps.

According to a BuzzFeed News report, Sensor Tower has operated at least 20 apps on iOS and Android to capture data, such as Free and Unlimited VPN, Luna VPN, Mobile Data and Adblock Focus. Some of those apps have been removed from the stores following BuzzFeed’s story.

I asked a lot of questions about Watchful’s source of data. “It’s all real users that give us access to this information. It’s all running on real devices, real users. We extract videos and screenshots from them,” Kahana said.

“It’s more like a panel of users that we have access to their devices. It’s not an SDK that is hidden in some app and collects information and do shady stuff,” he added.

You’ll have to trust him as the company didn’t want to elaborate further. Kahana also said that data is anonymized in order to remove all user information.

Images are then analyzed by a computer vision algorithm focused on differential analysis. The startup has a team in the Philippines that goes through all that data and annotates it. It is then sent to human analysts so that they can track apps and write reports.

Watchful shared one of those reports with TechCrunch earlier this year. Thanks to this process, the startup discovered that TikTok parent company ByteDance has been working on a deepfake maker. The feature was spotted in both TikTok and its Chinese sister app Douyin.

But Watchful’s customers aren’t news organizations. The company sells access to its service to big companies working in the mobile space. Kahana didn’t want to name them, but it said it is already working with “the biggest social network players and the biggest e-commerce players, mainly in the U.S.”

The startup sells annual contracts based on the number of apps that you want to track. It has raised a $ 3 million seed round led by Vertex Ventures .

Mobile – TechCrunch

Online reputation management: Seven steps to success

June 3, 2020 No Comments

30-second summary:

  • Apparently, it’s crucial to track your reputation in order to prevent PR crises. Moreover, monitoring your reputation enables you to discover valuable customer insights.
  • Founder and CMO at SEO PowerSuite and Awario, Aleh Barysevich, shares a strategy to tackle the challenges of online reputation management. 
  • Right from setting up your online reputation management (ORM) protocol to becoming proactive about getting reviews, there’s more to discover. 

There’s no need to explain the importance of reputation for businesses. The good word of your customers, potential or existing, is the best promotion tool you have. Meanwhile, scandals and criticism can ruin companies. Reputation becomes even more important during the times of crisis when emotions are heightened and any mistake can lead to a full-blown scandal. 

The internet really just amplified the importance of reputation: news and rumours travel fast, but on social media, they travel even faster (and reach more people). Who among us hasn’t checked the reviews before purchasing a product or hasn’t checked out a brand after seeing a friend praising it on social media?  No matter the size of your business, people are talking about you online, sharing their opinion on social media or leaving a review on Yelp and the likes of it.  

It would be wrong to think about online reputation as something separate from your “real-world” or offline reputation: with three billion social media users and counting your online reputation is simply your reputation, it affects purchasing decisions both online and offline.  

For example, Gillette’s polarizing campaign “The Best Men Can Be” gathered a lot of negative feedback (as well as some positive). Social media users publicly denounced the company and promised to stop buying Gillette razors. That wasn’t the first time a social media scandal led to calls for a boycott of the company, Nike and Uber being other notable examples.  

Besides the obvious need to track your reputation in order to prevent PR crises, monitoring your reputation enables you to discover valuable customer insights. Once you start paying close attention to your reviews and mentions online, you’ll learn what people love about your product, what they think you could improve, and what influences their decisions the most.  

All this makes reputation management more relevant than ever. You simply can’t ignore online conversations around your brand if you want to have a successful business. Luckily, the digital world gives us a lot more tactics and tools to monitor and actively improve our reputation than the offline world ever could. This article covers online reputation management step by step, giving you specific guidelines to follow.  

How to manage your reputation online 

Most businesses already conduct some type of online reputation management (ORM), for example, answering customers’ comments and posts where they were tagged. But to make your reputation crisis-proof, you need a robust workflow, and that’s what this article is all about. You can use these steps to revise your existing ORM workflow or build a new one from scratch. 

Step one: Set up your ORM protocol

Before you even start going through your online reviews, you need to establish some guidelines. These will help you and your team to know when to respond to reviews, do it appropriately and quickly, and know the best way to act if there’s a threat of a reputation crisis. This protocol can be as thorough as you like depending on how much you are synced with your team, but here are some questions to answer to figure out the guidelines: 

  • How fast should you answer? Obviously, the quicker your response, the better, but it’s a good practice to establish the minimum response time required for your team members. 
  • How transparent are you willing to be? This will help you determine if you want to go into all the nuances when responding to a customer or simply reassure them that you’re working on the issue. The recent trends prove that transparency is very much appreciated by customers.  
  • What tone of voice should you use? This, of course, will depend on your brand. Should you be cordial or professional and straight-to-the-point? Can you make jokes? Oftentimes a funny response to a complaint can go viral. For example, Oatly is one of the brands that heavily uses negative reviews in its marketing putting an ironic spin on them. But would it fit your brand’s image? 
  • Who will be the spokesperson(s) in case of a crisis? If the need arises, who will be giving the official statements on behalf of your brand? Is it the CEO, or the PR manager? Again, you can decide on the answer based on your company’s image — if you’re trying to build authentic relationships with customers and/or have a charismatic leader, it’s only logical that your CEO will do the talking.  
  • Should you automate your responses? Automation cuts your response time to seconds and allows you to save on staff, but can you be sure it won’t anger your customers? In the example below, the customer grew frustrated after trying to solve their problem on Twitter and getting the same scripted message from Amazon. 

  • Should you always respond? Some brands take their pride in the “always respond rule”, and for smaller brands, it’s actually a must — the more engagement you get, the higher your brand awareness, especially on social media. But once you start getting a ton of mentions at once, you might need to start prioritizing. Besides, sometimes negative reviews can just become trolling — and if there’s one rule you need to learn on the Internet, it’s “Don’t feed the trolls”.  

By answering the questions, you should have a clear outline of dos and don’ts for your social media, community, and PR managers. 

Step two: Choose and set up a monitoring tool

You could try tracking your online reviews and mentions manually, but without a specialized tool, it’s practically impossible. Online reputation management tools enable you to find mentions of your company on social media, in the news, and on review aggregator websites. There is an array of monitoring tools for different needs and budgets, such as AwarioBrandwatchReputology, and others. How do you choose the right one? 

As with the previous step, there are some questions you could ask yourself and your team to decide which tool to settle on: 

  • Platforms it covers. Of course, when we are doing online reputation management, the more feedback we find, the better. But for some businesses specific websites or social media platforms are crucial: for example, TripAdvisor for tourist guide companies or Instagram for clothing brands.  
  • Sentiment analysis. Sentiment analysis is one of the core features in ORM. It helps you focus on dealing with negative reviews first and see the overall share of negative and positive mentions of your brand.  
  • Special features. Are there any particular requirements that your team might have? Do you need influencer analytics to quickly prioritize reviews with the biggest reach first? Or do you maybe need a tool that could easily be integrated with your CRM/task manager?  
  • Pricing. Reputation management tools vary greatly in pricing going from Enterprise-level analytical powerhouses that cost thousands of dollars, to much more affordable options for mid- and small-sized businesses. Make sure the core features you’re looking for are available in the plan you’re ready to pay for! 

Most tools offer either free trials or demos to get you acquainted with them, so you can investigate before you are ready to invest. 

Once you settle on a tool, you need to set it up to start monitoring. You can monitor your brand name, the name of your products, the names of key public figures in your company. Don’t forget to include common misspellings of these words and phrases – it’s a common mistake that brands make when monitoring their reputation which results in missing a lot of feedback. 

setting online reputation management tool

Most tools allow you to choose some filters for your monitoring efforts: to find reviews only in a certain language, from certain countries or platforms.  

If there’s a topic that causes particular concern for your reputation (for example, Shell and oil spillage), you can create a separate monitoring alert for it using a Boolean search mode.  

Most reputation management tools have a notification settings tab where you can choose when and how you want to receive notifications. 

Now that you’re finished with your setup, it’s time to check your online reputation! 

Step three: Check sentiment analytics and mention spikes

The first thing to do every time you go to your online reputation management tool is to look at the dashboard. This is typically where all your analytics are visualized so you can notice if something is out of the ordinary right away.  

sentiment graph - online reputation management (ORM) tool

First, look at your overall sentiment and see the shares of positive and negative mentions. This will give you an understanding of your overall reputation. You can select different time frames to get a closer look at a certain moment in your company’s history, or vice versa choose as big of a time frame as you want to get a historical perspective.  

mentions graph

Other important graphs here are the number and reach of mentions, specifically, sudden spikes in it. A sudden spike in the number of mentions means that a lot of people are talking about you (hopefully, for a good reason) and a sudden spike in reach can also indicate that some influential account or website mentioned your brand. A lot of stories nowadays emerge on social media and paying attention to spikes allows you to get on top of the story right away.  

Step four: Deal with the social media mentions

Now that you’re sure that there are no reputation crises unfurling at the moment, it’s time to deal with individual mentions. I suggest focusing on social media first since it’s the media with the biggest “sense of urgency”, that is, the medium where people expect you to answer the quickest. According to the study by The Social Habit, 42% of social media users expect a brand to respond in 60 minutes or less.  

Most tools have some kind of a feed that gives you access to individual mentions. For now, filter out everything but the mentions from social media platforms: we will deal with the rest a bit later.  

Usually, social media mentions are sorted by date with the newest mentions displayed first. You can filter them to see negative mentions only to make sure you respond to the unhappy customers first, and then take a look at the neutral and positive mentions, thanking users and sharing favorable posts. Testimonials are an extremely powerful way to promote your brand, so don’t neglect the positive reviews you get, use them. 

The mentions, both positive and negative, can be a great source of customer insights as well. Pay attention to constructive feedback, you can even tag them to come back to them later or share them with your colleagues. 

Step five: Check review sites

Now that social media are dealt with, let’s move on to other types of review platforms: Google My Business, Yelp, TripAdvisor, Amazon, and any industry-specific platforms you might come across.  

To find these reviews, do the opposite of what we did in the previous step: filter out all the social media mentions as well as the news. Some tools like Awario even offer a whitelist feature which is used to prioritize certain domains — this could come in handy if you want to make sure you’re getting mentions from specific websites popular in your industry. 

Most websites allow you to respond to reviews once you verify your brand’s account — as with social media, start with the negative ones. You can also share the positive reviews on your website and social media through plugins or screenshots. 

If you allow reviews on your own website (if you’re running an eCommerce business, for example), now is a good time to go through them as well. 

Step six: Check mentions from the media 

Granted you haven’t noticed a sudden change in the sentiment or number of mentions at the third step, media outlets and blogs can wait until you’ve dealt with social reviews. Of course, if you’re in the midst of a PR scandal, the news becomes a much more important source. Also, these steps could be tackled by different teams – social and reviews sites can be taken care of by community managers and news and blogs can be handled by the PR professionals. Nevertheless, they will still be using the same online reputation management tool. 

Filter out everything but mentions from news and blogs. Then the workflow is pretty much the same: check the negative articles first, then the rest. You can reach out to bloggers and journalists to try to swing their opinion in case of negative coverage or thank them and possibly build lasting relationships.  

Step seven: Become proactive about getting reviews

Some communication specialists may treat reviews as a headache: the truth is people are much more likely to leave negative reviews than positive ones. This discrepancy can create a feeling of despair when it comes to online reputation management, but this only means that you need to become more proactive about getting reviews from your customers. 

The secret to getting more reviews is asking for them! You need to set up a consistent system of encouraging your customers to leave reviews on social media and review platforms. You can do it manually or use automation tools (Buffer, Mailchimp, Delighted) to schedule social media posts and emails encouraging users to leave reviews and add review-requests plugins to your website. Resharing positive reviews on social may also encourage other users to post praises to your business. You can even run a social media contest focusing on positive reviews as the main challenge for your followers. Get creative! 

To sum up

With proper preparations, online reputation management becomes a piece of cake. Once you have clear guidelines in place (which can be perfected over time) and set up a reputation management tool, there should be no trouble for you to make your reputation crisis-proof. 

Aleh Barysevich is Founder and CMO at SEO PowerSuite and Awario.

The post Online reputation management: Seven steps to success appeared first on Search Engine Watch.

Search Engine Watch

3 Ways To Build Customer Loyalty During A Crisis

June 3, 2020 No Comments

The 3 strategies in this blog are meant to help your brand better relate to customers during a crisis, ultimately increasing your customers’ lifetime value.

Read more at PPCHero.com
PPC Hero

Seven tips for full funnel SEO in 2020

June 2, 2020 No Comments

30-second summary:

  • Improving their organic search presence is the top inbound marketing priority for 61% of marketers.
  • It’s imperative that today’s marketers leverage paid search at every stage to create more sophisticated strategies because greater sophistication means less wasted budget and higher quality conversions. 
  • Erica Magnotto gives a crisp breakdown and categorization of how to make your sales and marketing funnel work through SEO.

Improving their organic search presence is the top inbound marketing priority for 61% of marketers. But many are unaware their tried-and-true search engine optimization tactics have lost their potency thanks to today’s more fluid marketing funnel.

Every marketer knows the marketing funnel: The famous upside-down triangle used to visualize the customer journey from “awareness” at the top to “action” at the bottom. Paid search is typically considered a lower-funnel tactic used to nudge customers toward a conversion. But in today’s digitally-dominated landscape, paid search plays a more integrated role. 

It’s imperative that today’s marketers leverage paid search at every stage to create more sophisticated strategies because greater sophistication means less wasted budget and higher quality conversions. Here are seven tips to plug in paid search throughout your customer relationships in 2020.

Phase one: Awareness

If you’re currently investing in awareness channels, you’re likely using a combination of programmatic display, video, social and influencers to connect with your audiencesBut don’t overlook paid search, which is also effective at driving new users to the website through competitor and educational campaigns.  

1. Competitor campaigns 

If a customer is looking for your direct competitor, it’s likely they’re in need of your services as well, so bidding on competitor terms is a great way to capture your competitor’s customers. Keep in mind though these keywords are usually expensive and receive lower quality scoresthey can help inform customers of their options within your industry. 

2. Brand education 

Customers looking to educate themselves on a particular product or service are likely to go to Google firstUse this knee-jerk reaction to send web traffic in your direction by adding specific content on your website that answer their questions. Blogs, white papers, FAQ pages, and industry updates are valuable forms of customer education that can develop brand awareness and promote trust with your audience. Even better, you can combine branding and lead generation by gating some of this content to collect user information that can later be repurposed for email marketing, retargeting, lookalike audiences and more. 

Phase two: Interest

After a user visits the website and gains familiarity with your brand keep your brand top of mind through retargeting list search ads and audience bidding.  

3. Retargeting List Search Ads (RLSAs)

RLSAs can help drive repeat visits to your website by directly targeting and bidding on those previous website visitors. These campaigns typically use tailored messaging, such as a discount, countdown or reminder to complete an action on your website, to create urgency. RLSAs can also be used with the brand and non-branded terms to entice user action in the decision phase of the funnel. 

4. Audience bidding

Marketers should consider applying in-market audiences to campaigns on observation mode to develop a clearer image of how each audience segment performs. Segments that perform or convert at a high level indicate interest from that grouping of potential customers and, therefore, are worth a higher investment through the use of bid modifiers. Bid up on these audiences to garner a stronger return. 

Phase Three: Decision

Invest in your brand terms to protect yourself from competitor interference once potential customers have made the conscious decision to engage with your brand. You can also implement extensions to impart more influence during the decision stage and garner increased user engagement. 

5. Bid on brand terms 

Brand terms serve two purposes in SEO: visibility and defence. First, it’s important to remain relevant to the SERPs by being represented in organic and paid results. Second, while you have the flexibility to bid on competitor terms, the competition can bid on your terms as wellBy creating a dedicated brand strategy for search, you can help cut down on competitors showing up in place of you.

6. Use extensions 

Extensions such as site links, callouts and structured snippets place more ads on SERPs, giving your ads more opportunity to influence in the decision of customersMarketers should use as many relevant extensions as possible to improve click-through-rates (CTRs) and higher quality scores. 

Phase four: Action

Search engines can optimize toward conversion action through automated bidding after users take action on your website through a paid ad. 

7. Automated bidding

Conversion or action data that are stored in Google or Microsoft Ads is repurposed to support automated bidding like Target Cost Per Action (CPA), Maximize Conversions or Target Return on Ad Spend (ROAS). These features support a variety of conversion goals through Google’s AI automation. Creating accurate conversions in your account is essential during the final step of the funnel.

Paid search is a foundational channel for driving lead generation initiatives, as it has vast capabilities. Search plays an integral role in every stage of the marketing funnel – not just at the top. Consider connecting with your in-house team or agency partners to revamp your paid search strategy for 2020 to ensure you are leveraging search across the funnel, and, ultimately, boosting the channels’ benefit to your bottom line.  

Erica Magnotto is Senior Search Engine Marketing Manager at R2i. 

The post Seven tips for full funnel SEO in 2020 appeared first on Search Engine Watch.

Search Engine Watch

Trump’s Antifa Obsession Is an Unconstitutional Distraction

June 2, 2020 No Comments

The president says the US will designate antifa a terrorist organization. Can he do that? Probably not.
Feed: All Latest

Atlassian launches new DevOps features

June 2, 2020 No Comments

Atlassian today launched a slew of DevOps-centric updates to a variety of its services, ranging from Bitbucket Cloud and Pipelines to Jira and others. While it’s quite a grabbag of announcements, the overall idea behind them is to make it easier for teams to collaborate across functions as companies adopt DevOps as their development practice of choice.

“I’ve seen a lot of these tech companies go through their agile and DevOps transformations over the years,” Tiffany To, the head of agile and DevOps solutions at Atlassian told me. “Everyone wants the benefits of DevOps, but — we know it — it gets complicated when we mix these teams together, we add all these tools. As we’ve talked with a lot of our users, for them to succeed in DevOps, they actually need a lot more than just the toolset. They have to enable the teams. And so that’s what a lot of these features are focused on.”

As To stressed, the company also worked with several ecosystem partners, for example, to extend the automation features in Jira Software Cloud, which can now also be triggered by commits and pull requests in GitHub, Gitlab and other code repositories that are integrated into Jira Software Cloud. “Now you get these really nice integrations for DevOps where we are enabling these developers to not spend time updating the issues,” To noted.

Indeed, a lot of the announcements focus on integrations with third-party tools. This, To said, is meant to allow Atlassian to meet developers where they are. If your code editor of choice is VS Code, for example, you can now try Atlassian’s now VS Code extension, which brings your task like from Jira Software Cloud to the editor, as well as a code review experience and CI/CD tracking from Bitbucket Pipelines.

Also new is the ‘Your Work’ dashboard in Bitbucket Cloud, which can now show you all of your assigned Jira issues. as well as Code Insights in Bitbucket Cloud. Code Insights features integrations with Mabl for test automation, Sentry for monitoring and Snyk for finding security vulnerabilities. These integrations were built on top of an open API, so teams can build their own integrations, too.

“There’s a really important trend to shift left. How do we remove the bugs and the security issues earlier in that dev cycle, because it costs more to fix it later,” said To. “You need to move that whole detection process much earlier in the software lifecycle.”

Jira Service Desk Cloud is getting a new Risk Management Engine that can score the risk of changes and auto-approve low-risk ones, as well as a new change management view to streamline the approval process.

Finally, there is a new Opsgenie and Bitbucket Cloud integration that centralizes alerts and promises to filter out the noise, as well as a nice incident investigation dashboard to help teams take a look at the last deployment that happened before the incident occurred.

“The reason why you need all these little features is that as you stitch together a very large number of tools […], there is just lots of these friction points,” said To. “And so there is this balance of, if you bought a single toolchain, all from one vendor, you would have fewer of these friction points, but then you don’t get to choose best of breed. Our mission is to enable you to pick the best tools because it’s not one-size-fits-all.”

Enterprise – TechCrunch

Job Search Engine Using Occupation Vectors

June 1, 2020 No Comments

I worked for the Courts of Delaware at Superior Court.

I started working there as the Assistant Criminal Deputy Prothonotary.

I changed positions after 7 years there, and I became a Mini/Micro Computer Network Administrator.

The Court used an old English title for that first position which meant that I supervised Court Clerks in the Criminal Department of the Court. In the second position, I never saw a mini/micro-computer but it was a much more technical position. I was reminded of those titles when writing this post.

What unusual job titles might you have held in the past?

A Job Search Engine Based on Occupation Vectors and a Job Identification Model

An Example of Job Search at Google:

job search example

For two weeks, Google was granted patents with the same name each of those 2 weeks. This is the first of the two patents during that period granted under the name “Search Engine.”

It is about a specific type of search engine. One that focuses upon a specific search vertical – A Job Search Engine.

The second patent granted under the name “Search Engine,” was one that focused upon indexing data related to applications on mobile devices. I wrote about it in the post A Native Application Vertical Search Engine at Google

The reason why I find it important to learn about and understand how these new “Search Engine” patents work is that they adopt some newer approaches to answering searches than some of the previous vertical search engines developed by Google. Understanding how they work may provide some ideas about how older searches at Google may have changed.

This Job Search Engine patent works with a job identification model to enhance job search by improving the quality of search results in response to a job search query.

We are told that the job identification model can identify relevant job postings that could otherwise go unnoticed by conventional algorithms due to inherent limitations of keyword-based searching. What implications does this have for organic search at Google that has focused upon keyword searches?

This job search may use methods in addition to conventional keyword-based searching. It uses an identification model that can identify relevant job postings which include job titles that do not match the keywords of a received job search query.

So, the patent tells us that in a query using the words “Patent Guru,” the job identification model may identify postings related to a:

  • “Patent Attorney”
  • “Intellectual Property Attorney”
  • “Attorney”
  • the like

The method behind job searching may include (remember the word “vector.” It is one I am seeing from Google a lot lately):

  • Defining a vector vocabulary
  • Defining an occupation taxonomy includings multiple different occupations
  • Obtaining multiple labeled training data items, wherein each labeled training data item is associated with at least:
    • (i) a job title
    • (ii) an occupation
  • Generating an occupation vector which includes a feature weight for each respective term in the vector vocabulary
  • Associating each respective occupation vector with an occupation in the occupation taxonomy based on the occupation of the labeled training data item used to generate the occupation vector
  • Receiving a search query that includes a string related to a characteristic of one or more potential job opportunities, generating a first vector based on the received query
  • Determining, for each respective occupation of the multiple occupations in the occupation taxonomy, a confidence score that is indicative of whether the query vector is correctly classified in the respective occupation
  • Selecting the particular occupation that is associated with the highest confidence score
  • Obtaining one or more job postings using the selected occupation
  • Providing the obtained job postings in a set of search results in response to the search query

These operations may include:

  • Receiving a search query that includes a string related to a characteristic of one or more job opportunities
  • Generating, based on the query, a query vector that includes a feature weight for each respective term in a predetermined vector vocabulary
  • Determining, for each respective occupation of the multiple occupations in the occupation taxonomy, a confidence score that is indicative of whether the query vector is correctly classified in the respective occupation
  • Selecting the particular occupation that is associated with the highest confidence score
  • Obtaining one or more job postings using the selected occupation, and providing the obtained job postings in a set of search results in response to the search query
  • Feature Weights for Terms in Vector Vocabularies

    It sounds like Google is trying to understand job position titles and how they may be connected, and developing a vector vocabulary, and build ontologies of related positions

    A feature weight may be based on:

    • A term frequency determined on several occurrences of each term in the job title of the training data item
    • An inverse occupation frequency that is determined based on many occupations in the occupation taxonomy where each respective term in the job title of the respective training data item is present.
    • An occupation derivative based on a density of each respective term in the job title of the respective training data item across each of the respective occupations in the occupation taxonomy
    • Both (i) a second value representing the inverse occupation frequency that is determined based, at least in part, on several occupations in the occupation taxonomy where each respective term in the job title of the respective training data item is present and (ii) a third value representing an occupation derivative that is based, at least in part, on a density of each respective term in the job title of the respective training data item across each of the respective occupations in the occupation taxonomy
    • A sum of (i) the second value representing the inverse occupation frequency, and (ii) one-third of the third value representing the occupation derivative

    The predetermined vector vocabulary may include terms that are present in training data items stored in a text corpus and terms that are not present in at least one training data item stored in the text corpus.

    This Job Search Engine Patent can be found at:

    Search engine
    Inventors: Ye Tian, Seyed Reza Mir Ghaderi, Xuejun Tao), Matthew Courtney, Pei-Chun Chen, and Christian Posse
    Assignee: Google LLC
    US Patent: 10,643,183
    Granted: May 5, 2020
    Filed: October 18, 2016


    Methods, systems, and apparatus, including computer programs encoded on storage devices, for performing a job opportunity search. In one aspect, a system includes a data processing apparatus, and a computer-readable storage device having stored thereon instructions that, when executed by the data processing apparatus, cause the data processing apparatus to perform operations.

    The operations include defining a vector vocabulary, defining an occupation taxonomy that includes multiple different occupations, obtaining multiple labeled training data items, wherein each labeled training data item is associated with at least (i) a job title, and (ii) an occupation, generating, for each of the respective labeled training data items, an occupation vector that includes a feature weight for each respective term in the vector vocabulary and associating each respective occupation vector with an occupation in the occupation taxonomy based on the occupation of the labeled training data item used to generate the occupation vector.

    The Job Identification Model

    Job identification model

    Job postings from many different sources may be related to one or more occupations.

    An occupation may include a particular category that encompasses one or more job titles that describe the same profession.

    Two or more of the obtained job postings may be related to the same, or substantially similar, occupation while using different terminology to describe a job title for each of the two or more particular job postings.

    Such differences in the terminology used to describe a particular job title of a job posting may arise for a variety of different reasons:

    • Different people from different employers draft each respective job posting
    • Unique job titles may be based on the culture of the employer’s company, the employer’s marketing strategy, or the like

    occupation taxonomy

    How an Job Identification Model May Work

    An example:

    1. At a first hair salon marketed as a rugged barbershop, advertises a job posting for a “barber”
    2. At a second hair salon marketed as a trendy beauty salon, advertises a job posting for a “stylist”
    3. At both, the job posting seeks a person for the occupation of a “hairdresser” who cuts and styles hair
    4. In a search system limited to keyword-based searching, a searcher seeking job opportunities for a “hairdresser” searchings for job opportunities using the term “barber” may not receive available job postings for a “stylist,” “hairdresser,” or the like if those job postings do not include the term “barber”
    5. The process in this patent uses a job identification model seeking to address this problem

    The job occupation model includes:

    • A classification unit
    • An occupation taxonomy

    The occupation taxonomy associates known job titles from existing job posts with one or more particular occupations.

    During training, the job identification model associates each occupation vector that was generated for an obtained job posting with an occupation in the occupation taxonomy.

    The classification unit may receive the search query and generate a query vector.

    The classification unit may access the occupation taxonomy and calculate, for each particular occupation in the occupation taxonomy, a confidence score that is indicative of the likelihood that the query vector is properly classified into each particular occupation of the multiple occupations in the occupation taxonomy.

    Then, the classification unit may select the occupation associated with the highest confidence score as the occupation that is related to the query vector and provide the selected occupation to the job identification model.

    An Example of a Search Under this Job Opportunities Search Engine:

    1. A searcher queries “Software Guru” into a search box
    2. The search query may be received by the job identification model
    3. The job identification model provides an input to the classification unit including the query
    4. The classification unit generates a query vector
    5. The classification unit analyzes the query vector given the one or more occupation vectors that were generated and associated with each particular occupation in the occupation taxonomy such as occupation vectors
    6. The classification unit may then determine that the query vector is associated with a particular occupation based on a calculated confidence score, and select the particular occupation
    7. The job identification model may receive the particular occupation from the classification unit
    8. Alternatively, or besides, the output from the classification unit may include a confidence score that indicates the likelihood that the query vector is related to the occupation output by the occupation taxonomy
    9. The occupation output from the occupation taxonomy can be used to retrieve relevant job postings
    10. Specifically, given the output of a particular occupation, the job identification model can retrieve one or more job postings using a job posting index that stores references to job postings based on occupation type

    11. The references to job postings that were identified using the job posting index are returned to the user device
    12. The obtained references to job postings may be displayed on the graphical user interface
    13. The obtained references to job postings may be presented as search results and include references to job postings for a “Senior Programmer,” a “Software Engineer,” a “Software Ninja,” or the like
    14. The job postings included in the search results were determined to be responsive to the search query “Software Guru” based at least in part on the vector analysis of the query vector and one or more occupation vectors used to train the occupation taxonomy and not merely based on keyword searching alone

    Takeaways About this Job Search Engine

    In addition to the details about, the patent tells us how an occupation taxonomy may be trained, using training data. It also provides more details about the Job identification model. And then tells us about how a job search is performed using that job identification model.

    I mentioned above that this job search engine patent and the application search engine patent are using methods that we may see in other search verticals at Google. I have written about one approach that could be used in Organic search in the post Google Using Website Representation Vectors to Classify with Expertise and Authority

    Another one of those may involve image searching at Google. I’ve written about Google Image Search Labels Becoming More Semantic?

    I will be posting more soon about how Google Image search is using neural networks to categorize and cluster Images to return in search results.

    Copyright © 2020 SEO by the Sea ⚓. This Feed is for personal non-commercial use only. If you are not reading this material in your news aggregator, the site you are looking at may be guilty of copyright infringement. Please contact SEO by the Sea, so we can take appropriate action immediately.
    Plugin by Taragana

    The post Job Search Engine Using Occupation Vectors appeared first on SEO by the Sea ⚓.

    SEO by the Sea ⚓

    Scalable Analytics Team Structures: Centralized vs Decentralized

    June 1, 2020 No Comments

    The structure of your analytics team should be scalable. This blog defines centralized and decentralized analytics teams and looks at the pros/cons of each.

    Read more at PPCHero.com
    PPC Hero

    6 VCs share their bets on the future of work

    June 1, 2020 No Comments

    As tech companies like Twitter and Facebook gear up for longer-term remote work solutions, the future of work is becoming one of the more exciting opportunities in venture capital, Charles River Ventures general partner Saar Gur told TechCrunch.

    And as loneliness mounts with shelter-in-place orders implemented in various forms across the world, investors are looking for products and services that foster true connection among a distributed workforce, as well as a distributed society.

    But the future of work doesn’t just entail spinning up home offices. It also involves gig workers, freelancers, hiring tools, tools for workplace organizing and automation. The last couple of years have particularly brought tech organizing to the forefront. Whether it was the Google walkout in 2018 or gig workers’ ongoing actions against companies like Uber, Lyft and Instacart for better pay and protections, there are many opportunities to help workers better organize and achieve their goals.

    Below, we’ve gathered insights from:

    Saar Gur, Charles River Ventures 

    What are you most excited about in the future of work?

    Future of work is one of the most exciting opportunities in venture.  

    Pre-COVID, few tech companies were fully remote. While it seems obvious in retrospect, the building blocks for fully remote technology companies now exist (e.g. high-speed internet, SaaS and the cloud, reliable video streaming, real-time documents, etc.). And while SIP may be temporary, we feel the TAM of fully remote companies will grow significantly and produce a number of exciting investment opportunities.

    I don’t think we have fully grokked what it means to run a company digitally. Today, most processes like interviewing, meetings and performance/activity tracking still live in the world of atoms versus bits. As an example, imagine every meeting is recorded, transcribed and searchable — how would that transform how we work?   

    There is an opportunity to re-imagine how we work. And we are excited about products that solve meaningful problems in the areas of productivity, brainstorming, communication tools, workflows and more. We also see a lot of potential in infrastructure required to facilitate remote and global teams.

    We are also excited by companies that are enabling new types of work. Companies like Etsy (founded 2005), Shopify (2004), TaskTabbit (2008), Uber (2009), DoorDash (2013) and Patreon (2013) have helped create a new workforce of entrepreneurs. But many of these companies are over a decade old and we fully expect a new wave of companies that give more power to the individual.

    Startups – TechCrunch