RSS
 

Archive for the ‘Searching’ Category

Hashtag Database

23 Mar 2014

hashtags on products - DaveTavres.comIn 2010 when I first wrote about QR codes, several friends say, ‘nah, they won’t take off’. I’m sure people said the same thing about hashtags. Well, hashtags will soon be included on packaging for major brand products, just like QR codes.

#Hashtags are more than common now and there are dozens of websites that promise to give businesses ‘tracking data’ to determine how popular a tag is – in other words, see just how many people passed along the message that was trying to be conveyed. Facebook and Twitter are obviously the most popular social media outlets to use them, but more and more television stations, individual TV shows, movies and of McDonald's hashtag - DaveTavres.comcourse whatever the internet fad of the moment are using them.

Quick quiz… you write a quick post online about wanting to watch the Olympics. Do you write:
A) “Can’t wait to watch the #Olympics tonight!”
B) “Can’t wait to watch the #SochiOlympics tonight!”
C) “Can’t wait to watch the #2014Olympics tonight!”

Of course the answer is – all are correct – but if you want to ‘correctly’ attribute the tag so that searches will take you to the Olympic organizer’s website, which is the “most” correct?

Here’s another one – what if you’re writing about the bone disease Avascular Necrosis which is often shortened to AVN and you add “#AVN” to the post, but don’t check it before you post it… then find out that AVN is probably far more commonly known as the “Adult Video News” – which runs their own version of the #Oscars for porn. Whoops.

Disney hashtag - DaveTavres.comOne last scenario… you’re working on article about camping and travelling and you sidetrack a little about visiting a nearby rainforest. Your article will be released to a broad distribution newsletter that will also be printed and emailed and posted online. You want to ’correctly’ give reference to an organization that raises funds for “saving the rainforests”. You’re not an expert on rainforests and you didn’t easily find a tag for it… So which tag do you use?

You logically use “#SaveTheRainforests”. Then you find out that an oil conglomerate is using that tag in their advertisements, trying to show how environmental they are – when in fact, they clear cut thousands of acres of rainforests a year but ‘save’ 1% of each forest. You may not want to support that company… but now you are supporting their message. Another “whoops”.

Facebook hashtag - DaveTavres.comSo, where do you find the tag that GreenPeace uses when they talk about saving a rainforest? In this article title “The Great Bear Rainforest Victory” or this one “Protecting forests”? Nope. They don’t have any hashtags there. So what should you use?

A website idea

2014 Volkswagen Game Day Commercial: Wings

Okay… how about a website that lists all the hashtags that a company uses for their various advertising campaigns and brands? On the same site you can search (and watch) that fun commercial where the car guys get their wings… Was the tag #Wings or #VW100000… no, it was just #VW (kind of a missed opportunity).

The revenue model would be purely ad-based. This isn’t likely to be a popular consumer website (but who really knows?) If it caught on, companies could pay for click traffic when someone searches for a specific tag to find the ‘official’ company that started using the tag.

In Closing

As hashtags become more commercial, brands will be using them more. Not just in commercials, but in more printed materials, including product packaging. So why not get a jump on the “hashtag database” now? HashBase.com?

 

A new search box trend

31 Dec 2013

Recently I’ve seen several websites that are ‘hiding’ the site search functionality. I don’t understand why… and I don’t like it. It’s hard enough to find what you’re looking for in a lot of modern navigation bars. Now with just a search ‘icon’, it makes it even harder to find. Here’s two of the many examples I’ve seen lately:

Rally Software
Rally search - DaveTavres.com

DealNews.com
DealNews search - DaveTavres.com

I think it’s much easier to quickly scan for that search text field – everyone is use to it, it’s easier to find than a single icon that could be lost with other icons and it can just as easily be reduced to save space but automatically expand when a user clicks into the field.

Facebook is certainly not winning award for ‘best in user experience’, but one thing that hasn’t change is the search box. It’s there… small and large.

Facebook 2007 vs. 2013
Facebook 2007 - DaveTavres.comFacebook 2013 - DaveTavres.com

Some things just shouldn’t change…

 

Photo search

16 Jan 2012

Google Image SearchI collect FAR too many photos from the web… stuff I find interesting or of my favorites topics like Bodie, Walt Disney’s apartment, the Disneyland Railroad, history or movie and TV filming locations. Sometimes I get an email or see a photo that has no description or information, which used to make it nearly impossible to figure out what the photo is, or where it was taken.

Well, Google has done it again. Their "Search by image" feature lets you find that image, or similar ones, without having to type in a huge description and guessing what might be useful in tracking down some information.

Here’s how it works:

  1. Save that amazing photo from your email or a website to your desktop. For this example, save the picture below (click the image so it opens the photo, then right-click and choose ‘Save picture/image as’).
    train                             Where was this train photo taken?!
  2. In your favorite browser, go to www.Google.com/images. You can also just go to Google and click the "Images" link in the top navigation.
  3. Move the browser window to the side so you can see the image icon that you saved on your desktop. Then, drag the image icon into the search field.
    The search field will change to a "Drop image here" box – just drop… drop the image there.
    Google Image Search
  4. Google will upload that image, analyze it and give you a best guess of what the original image was, as well as give you visually similar images.
    Google Image Search

From here, it makes things MUCH easier to figure out if that photo that claims a man eating shark was caught off the coast of Seattle is real, or where that spectacular photo of two steam locomotives crossing a trestle next to the ocean was taken.

 

Block domains from Google search results

11 Jan 2012

GoogleIf you have a Google / Gmail account and stay logged in, Google has provided a fantastic tool for blocking those websites or domains that never return anything good or trustworthy.

Just follow these steps:

  1. Go to www.Google.com.
  2. Click the "Sign in" button at the top-right corner and sign in with your account.
  3. Go to http://www.google.com/reviews/t?hl=en
  4. In the first box, enter: http://wikipedia.org
    In the second box, enter your reason for blocking the site.
    Click the "Block site" button.GoogleBlocking

Within a few hours no Wikipedia results will show up in your Google search results. Sadly, this only works if you are logged into your Google/Gmail account.

 

Job matching business idea…

01 Dec 2011

Job matching business idea… www.DaveTavres.comBefore the mid-1990s people use to search for jobs by opening a newspaper, going to an unemployment office, talking with friends and associates or just walking into a company and asking if they were hiring. That process was still considered normal until the early 2000s, even though Monster.com, HotJobs.com, CareerBuilder.com and dozens of other job search engines came into existence.

For a few short years newspapers moved their classified jobs sections to their websites, but it became so cumbersome that even they turned over that work to the big job search websites. Today, you’d be hard pressed to find a single large-scale newspaper who runs its own classified jobs section.

The advent of “powerful” job search functionality has been an amazing tool. Job seekers could whittle down their search into manageable chunks without having to sift through dozens (or hundreds) of jobs that didn’t fit their qualifications. The ability to search for jobs in a specific industry, within a specific geographic location and by a general salary range is now commonplace, but it’s still incredibly inefficient and frustrating – on ALL of the job search engines.

CareerBuilder.comDice.comHotJobs.comIndeed.comtechies.comMonster.com

HR and hiring managers count on being able to search resumes by keywords, rather than having to read through dozens (or hundreds… or thousands) of applicants. There’s an entire industry dedicated to re-writing your just to get noticed during those automated searches. In this new environment, HR and hiring managers only ever READ a handful of actual resumes when they are filling a position.

So, here’s a new idea that isn’t currently out there…

An AI (artificial intelligence) job matching tool that tells the job seeker which jobs they should apply to, rather than just applying to everything, hoping someone will see their information.

The current system works like this:

  1. A company writes a job posting and puts it online.
  2. A job seeker finds dozens of postings within a search criteria.
  3. The job seeker tweaks their resume, then submits it, and usually never hears back from the company (mainly because that company got 2,853 resumes for that one position.)
  4. The job poster does a keyword search on the respondents and ends up with a few dozen resumes. They throw out the ones that aren’t formatted well or are difficult to understand at a glance. They get down to a handful of applicants and then reads those resumes.

BottomlessPit
bottomless pit

The issues:

  • Job seekers spend an enormous amount of time customizing and responding to job postings, even though their submission will never been seen by a person.
  • Job posters get far too many applicants to be able to give any time to all of them.

A possible solution:

  • The job seeker creates an account and fills in some basic information, i.e. location desired, salary range (lower AND upper), job type (part-time, full-time, etc.)
  • The job seeker uploads their resume to the AI website.
    • The AI website:
      • strips out all of the useless words (and, if, the, is, etc.).
      • catalogs all of the words and phrases, building an index of the words and their proximity to each other.
      • scores and ranks each word and phrase individually, as well as scoring and ranking the word proximities.
  • The AI engine then queries it’s database of jobs based on the simplified “basic information” the job seeker entered AND on the overall score for that resume.
  • The result is a list of very few job postings that the job seeker is qualified for, thus, more likely to be seen by a human being and reducing the amount of time the job seeker spends submitting resumes to jobs they won’t get. Plus, job seekers can upload different versions of their resume to be scored and for a different set of job results.

Obviously the database of positions is key.

Jobs database - www.DaveTavres.comIndeed.com is a large job search website that pulls in job posting from dozens of other job search websites. As an aggregator, they have tried to make it easy to search for “all” jobs on the web at once, rather than having to create an account on every job search website out there. The beauty is that Indeed.com has an API / web service for searching the jobs in their database.

The AI website can download each job posting from Indeed.com every day, scoring and ranking each post in the same way it does for uploaded resumes. This is how the proposed system now becomes a job matching tool, rather than a job search engine.

Job Matching IS the future. Whether the economy is good or bad, it always comes down to efficiency and money. People don’t want to spend hours and hours every day trying to find, then re-write their resumes to match a job, just to submit it and never get a reply. Just as companies don’t want to pay more people than they need to find candidates for open positions.

There’s a lot more to this idea, but this is plenty without posting an entire specification online. I’ve had this idea for a while, but I don’t have the money to execute, so someone else out there will have to do it. I predict that in the next five years, this will be a standard tool on all job search engines.