pancakes

MicrostockGroup Sponsors


Author Topic: Is what you are currently doing working?  (Read 17761 times)

0 Members and 1 Guest are viewing this topic.

« on: October 03, 2017, 16:46 »
+1
Are you happy with the current direction of the industry?  What needs to change?  Where do you see the future?  What do you consider our industries biggest problems? 

Many of you know who I am, and what I am doing.  I have posted here before about what I do to try and solve some of the industries problems.  I am working on perfecting a better, smarter, learning, and predictive image search.  I have also opened my past agencies enterprise-level sales and delivery platform to photographers, giving 100% of the license fee they set, for a flat fee to run the system.

I wanted to comment on some trends am following, many resulting in my current direction, I hope to receive your constructive feedback, and see what we can do together to make things better for all of us. 

Content creation in our industry is growing exponentially.  Now that just about everyone with a camera in his or her pocket, anyone can create content for our industry.  Yes, I know it takes more than a smart camera to take great photos. However, the barriers to entry into the stock photo market space continue to decline.  Image quality continues to improve, and finding the best (most relevant) images to represent a company, product, brand or idea will continue to vex image buyers.  Manual image curation is not enough, and cannot possibly predict the future market need.  I built my first stock photo agency Corner House Stock Photo on the inadequacies of stock photo agency curation.  I took on released houses and home lifestyle images that other photo agencies turned away; we filled a niche others did not see.  I am an architecture shooter and had direct insight into what real buyers in the space wanted and were looking for, which the other agencies did not see. 

With this influx of new stock photo content, oversupply and competition, agencies have even less of an incentive to take on additional out of scope images that may or may not sell.  It appears that the new" plan or trend, is to allow buyers to request content on demand.  On-demand content is nothing new, look up the history of On Request Images and others following the same model like ImageBrief.  Yes, we can hear it straight from the horse's mouth what buyers are "looking for" when they put out a request or brief.  However shooting such content with a single consumer's particular product or service in mind, on speck is risky business for the image creator, locking up their time on creating images for one buyer.  Yes, the request can be expanded upon and shot for other subjects. The short version - In my opinion retrying old failed business models, is NOT a good "new" plan for the future.

My focus has been on a smarter, learning and predictive image search.  A search that is all-inclusive, and lets the image buyer decide what they want to see and we simply show them more of what they want and less of what they don't.  The search engine model works, we use it every day for other industries, why not stock photos?   It is time to embrace the technology that is pushing our industry into the future.  Use machine learning and artificial intelligence to help buyers find images and help photographers get images licensed.  I am the first to say that my system is not perfect, we have had our fair share of growing pains, but it is getting better, faster and smarter every day.  With enough data, we can predict cycles and trends sharing this information with image creators.  It all depends on how you use the massive amount of data collected and available.  I am actively working on this. 

Let me know where you see the future.  Is this a problem that can or needs to be solved?  How?  Do you think search and data derived from search can help us or are we better off in our current path guessing and shooting in the dark? I am open to your constructive thoughts.


« Reply #1 on: October 04, 2017, 00:24 »
+2
I agree I think the next step forward is a really smart search engine if anyone can deliver that then they will have a serious advantage

« Reply #2 on: October 04, 2017, 01:36 »
+3
The problem is that most people here would define a "smart search engine" as an engine that finds THEIR clips.

« Reply #3 on: October 04, 2017, 01:43 »
0
The problem is that most people here would define a "smart search engine" as an engine that finds THEIR clips.
Indeed! It needs a lot of input from buyers to be meaningful and not what the agencies think is best for them. Then my masterworks would obviously naturally find their way to the top ;-).

« Reply #4 on: October 04, 2017, 01:58 »
+5
Sorry. No idea what you do....beyond this email....

dpimborough

« Reply #5 on: October 04, 2017, 03:28 »
+4
"Content creation in our industry is growing exponentially.  Now that just about everyone with a camera in his or her pocket, anyone can create content for our industry. "

No

The content being submitted is rubbish on the whole, poor quality, poorly keyworded.

Most of the contributions should never be accepted in the first place.

What this industry needs is better quality control.

« Reply #6 on: October 04, 2017, 03:44 »
0
"Content creation in our industry is growing exponentially.  Now that just about everyone with a camera in his or her pocket, anyone can create content for our industry. "

No

The content being submitted is rubbish on the whole, poor quality, poorly keyworded.

Most of the contributions should never be accepted in the first place.

What this industry needs is better quality control.
Better quality control costs that's why if a search engine can do it it would be a huge financial edge. The agencies aren't going to pay for QC unless customers pay a premium which is why the more exclusive agencies have a niche.

« Reply #7 on: October 04, 2017, 10:32 »
+1
Better quality control costs that's why if a search engine can do it it would be a huge financial edge. The agencies aren't going to pay for QC unless customers pay a premium which is why the more exclusive agencies have a niche.

Shutterstock is neither exclusive or niche,  yet they have no problem in attracting premium buyers.  Their enterprise customers contribute to almost  a third of Shutterstock's turnover, a colossal amount, and if they cannot find what they want the enterprise team will find it for them even delving down to where our images are found, that's where your SODs come from.  I guess that the more difficult it was to search on Shutterstock the easier it was to convert buyers to their enterprise platform, and hence the situation we have now.

If quality contol was counter productive for Shutterstock there will surely come a point when that ceases to be the case and a new method of searching will be required.  Searches dependant on keywords, tittles and descriptions alone will never be good enough on a site like Shutterstock, too much corupted data, but if you combine it with image recognition you can eliminate many inappropriately keyworded images in the search.  Eventually, image recognition will become the dominant factor in searches and keywords will become secondary.





ShadySue

  • There is a crack in everything
« Reply #8 on: October 04, 2017, 10:40 »
0
I agree I think the next step forward is a really smart search engine if anyone can deliver that then they will have a serious advantage
Provided that in parallel they can manage to stop spam from contributors.


« Reply #10 on: October 04, 2017, 16:22 »
0
A review as we have evolved over the years.  Yes, I have been working on this for many years, because I believe this is the direction the industry is heading.  My goal is still the same. 

We are the search engine and marketplace for the stock photo industry with a focus on smart image search and discovery. 

Including every stock photo in the industry within our search results is our goal.  PicturEngine is not another agency or aggregator and does not take any commissions.  I am a photographer and agency owner myself (picturesque.com and cornerhousestock.com).  After running my other stock photo agencies for many years, I realized our industry has evolved and does not need another stock photo agency or new pricing model.  We need a better, smarter way to search for images.

PicturEngine operates on the traditional search engine model with organic and paid search results.  The best, most liked, viewed and clicked on images float to the top of the search results, while at the same time, our smart machine learning search allows many images not previously or regularly seen by art buyers to be discovered.  We use the art buyer's own searching patterns and buying habits to suggest unique, never before seen images helping buyers find that perfect image for their needs.  It is not always about the most popular or most "used" image. 

PicturEngine does not compare image prices.  It is our goal to connect image buyers as close to the source creator (or agency that pays the creator directly) as possible, thus reducing the unnecessary intermediaries that exist in our industry.  In theory, this will give the creator more money to keep shooting and save the buyer some time searching and money too.  Although, some photographers have the same image listed for $1 at agencies and list on our platform for $25, so it is not perfect.  Providing creators with good usable data helps them make informed decisions.

Agencies over 1 million images make up our organic content, (this gives image buyers a compelling reason to come to search us first) smaller agencies and individual creators have the option pay a small advertising fee (or use our platform) to be suggested or injected into our smart search.  Think of it like shuffling a deck of cards, organic on the one hand and paid on the other; our suggestion engine decides the placement and ranking of images for that particular buyer's needs. 

Individual creators already using a licensing platform can utilize our Advertising Only system.  Several existing platforms like PhotoShelter and PhotoDeck are fully automated; we push images into the database immediately upon signup to PicutrEngine.  Other platforms use a CSV or XML upload of a "data file" (no need to upload images directly for Advertising Only) until we bring these other platforms into the automated system.  It does take a couple of weeks for PicturEngine to create a visual fingerprint and for images to begin gaining a proper ranking in our system.  We also remove exact duplicates from agencies when creators use our system.  We look at the whole picture, all of its data and visual content for ranking (along with what image buyers think of the image per the terms searched), not just the existing attached captions and keywords.  The industry has a keywording problem that we are actively addressing with our smart search.

PicturEngine has a direct to seller marketplace called the Photographer Platform.  We make image licensing directly from image creators simple for photographers on our platform.  Photographers maintain 100% control of their images, set their pricing and keep 100% of the image license fees.  We have changed the platform system pricing recently offering tiers as well as our original unlimited plan.

Last but certainly not least, we provide many other valuable services such as image keywording, industry pricing analysis, analytics, and performance metrics.  We will be providing data and analytics detailing how your images are performing and how to edit better, shoot, organize, caption, and keyword your images for better performance within our system and yours.  We want to provide the data on where the industry has been, where it is today and where it is going. 

PicturEngine is continuing to evolve.  I want us to be the best image search and discovery engine possible.  Helping this industry that I have been a part of for over 20 years.  The industry cannot continue its current course; something has to change.  I have asked for help during our beta and sincerely appreciate all of those that helped us develop into what we are today.  Data science is an expensive, long and tedious process and involves a lot of testing.  We are not perfect, that is why I continue to ask how we can be better.  If you have a suggestion or complaint, bring it to me, personally, and I will address it.  My goal is like yours to make this industry sustainable in the future.

Feel free to ask questions; we are always here to help.  Also please see our FAQs for any questions we have answered.
PicturEngine
http://www.picturengine.com/
Pricing http://www.picturengine.com/#pricing-plans
FAQs http://support.picturengine.com/

Again, my goal in posting here is to get your feedback on what is working for you.  What is not working and where you see the future of our industry.  I am here to try and help.
« Last Edit: October 07, 2017, 20:42 by PicturEngine-JustinB »

« Reply #11 on: October 04, 2017, 16:49 »
+2
I agree I think the next step forward is a really smart search engine if anyone can deliver that then they will have a serious advantage
Provided that in parallel they can manage to stop spam from contributors.
Not so much stop as render it counter  productive by burying content inappropriately described.

« Reply #12 on: October 04, 2017, 16:53 »
0
Better quality control costs that's why if a search engine can do it it would be a huge financial edge. The agencies aren't going to pay for QC unless customers pay a premium which is why the more exclusive agencies have a niche.

Shutterstock is neither exclusive or niche,  yet they have no problem in attracting premium buyers.  Their enterprise customers contribute to almost  a third of Shutterstock's turnover, a colossal amount, and if they cannot find what they want the enterprise team will find it for them even delving down to where our images are found, that's where your SODs come from.  I guess that the more difficult it was to search on Shutterstock the easier it was to convert buyers to their enterprise platform, and hence the situation we have now.


Quite, so customers are also paying a premium via the enterprise team who are in effect undertaking post acceptance QC or a filter. If the search engine is clever enough to reduce/eliminate that need then thats a big opportunity.

« Reply #13 on: October 04, 2017, 17:39 »
+5
but content can also rise by being paid for, thereby bypassing the very thing you are trying to promote - the best images in front of buyers.

namussi

« Reply #14 on: October 04, 2017, 20:29 »
+6
PicturEngine-JustinB

Your posts are far too long. And the important bits are buried a long way down in the text.


« Last Edit: October 04, 2017, 20:41 by namussi »

namussi

« Reply #15 on: October 05, 2017, 01:47 »
0
PicturEngine-JustinB

Your posts are far too long. And the important bits are buried a long way down in the text.

My tone seemed a bit hostile there. Apologies. Wasn't meant to.

« Reply #16 on: October 05, 2017, 02:03 »
+4
The proof is in the eating of pudding though......if I search "aardvark" I get lots of pics of birds and other random animals......


« Reply #17 on: October 05, 2017, 02:13 »
0
The proof is in the eating of pudding though......if I search "aardvark" I get lots of pics of birds and other random animals......

Yes, the industry has a keywording problem. And I don't want to find this image when I search for "tiger cat", which is neither a cat nor a tiger... It's a sand tiger shark / grey nurse shark, and of course shouldn't have "cat" or "feline predator" in the tags. Also another reason why automated tagging systems are completely idiotic (I assume this is why "feline predator" is in there, because "tiger" was there.

And as soon as paid content is included, well, that's pretty obvious... Reviewers not looking at keywords is the problem with unwanted results.

A true smart engine would require ACTUAL people going through thousands and thousands of searches and removing unwanted results that are false matches. That would be a capping system I could get behind.  :D
« Last Edit: October 05, 2017, 02:28 by increasingdifficulty »

« Reply #18 on: October 05, 2017, 02:20 »
+1
The proof is in the eating of pudding though......if I search "aardvark" I get lots of pics of birds and other random animals......

Yes, the industry has a keywording problem. And I don't want to find this image when I search for "tiger cat", which is neither a cat or a tiger for that matter... It's a sand tiger shark / grey nurse shark, and of course shouldn't have "cat" or "feline predator" in the tags. And as soon as paid content is included, well, that's pretty obvious... Reviewers not looking at keywords is the problem with unwanted results.

A true smart engine would require ACTUAL people going through thousands and thousands of searches and removing unwanted results that are false matches. That would be a capping system I could get behind.  :D
The dream is something that achieves this without humans....whether thats a practical possibility this side of the launch of the SS Enterprise I'm not knowledgable enough to judge

« Reply #19 on: October 05, 2017, 02:22 »
0
The dream is something that achieves this without humans....whether thats a practical possibility this side of the launch of the SS Enterprise I'm not knowledgable enough to judge

Yes, maybe in a 100 years.

« Reply #20 on: October 05, 2017, 03:06 »
+3
As a video contributor AND a production company owner - I can see both sides. In the world of video we find ourselves searching for a specific subject in all the agencies with poor results.
A search for "Business woman" yields 222,817 results on Shutterstock, while looking for a hot, relevant issue like "Islamic terror" gives you just over 1,500 clips, many of them irrelevant, and 70% editorial.
What I see as the challenge is finding these places where demand is high and supply is low (in the video market these places still exist I believe)
What I need as a contributor is a search tool that finds the gaps between high volume searches and low volume sales, or something like that...

« Reply #21 on: October 05, 2017, 04:16 »
+1
Thanks Jo Ann Snover. I have also looked at the site.
I remain confused.

This site is a bit like Google Images which also links to my pics (and then - hopefully - an Agency). (Lots of my pics are already on Picturengine through assorted agencies and I have paid nothing.)

But if I want to include my own website (set my own prices etc) then (unlike Google Images) it will only find me if I pay a fee. Is that right?

I have no idea what Sales and Delivery means if it is a search engine.

I'm sure I'm missing something here.....









niktol

« Reply #22 on: October 05, 2017, 07:51 »
0

This site is a bit like Google Images which also links to my pics (and then - hopefully - an Agency). (Lots of my pics are already on Picturengine through assorted agencies and I have paid nothing.)

But if I want to include my own website (set my own prices etc) then (unlike Google Images) it will only find me if I pay a fee. Is that right?



Pretty much. Except there is no way of knowing how many buyers - that is people who actually do open their wallet - are using this engine (unlike Google images), therefore its usefulness for artists remains questionable. So is its"smartness", because there is no info on which features are included in the learning algorithm. It could be either very basic or unnecessary complicated with a potential of throwing out lotsa potentially useful stuff.

I have no idea what Sales and Delivery means if it is a search engine.

I'm sure I'm missing something here.....


There is just too much irrelevant info in the posts. Perhaps it's something the poster did before but does not do anymore. The mentioned picturesque.com and cornerhousestock.com link to www.picturengine.com/.

« Reply #23 on: October 05, 2017, 13:54 »
0
The proof is in the eating of pudding though......if I search "aardvark" I get lots of pics of birds and other random animals......

Yup, I just bought one of "Bottle of Juice" surrounded by coffees and teas.

« Reply #24 on: October 05, 2017, 15:53 »
0
Sorry for the long responses.  I feel it is often useful to understand my thinking; this hopefully gives you a better understanding of what we are doing and why. 

I am creating a better, smarter, learning image search.
We are a search engine and marketplace. We have to pay for the smart search somehow, so we have broken it up into parts to benefit creators.   

Search engines traditionally operate on advertising revenue.  Take a look at Google, (NOT Google images) paid vs. organic search results.  (Google Images does not make Google money directly YET, but that is an entirely different and long topic.)

The current search engine PPC (Pay Per Click) advertising models do not work to benefit smaller creators in our industry.  These models reward those with the most money to advertise for paid search or the sites with the best SEO for organic traffic (usually the ones with the money to pay for PPC.)  On Google, advertising clicks for searches such as "Insurance" pay as much as $55 per click!  Individual contributors to the stock photo industry cannot support these high PPC prices and would not/do not use it.  In Google, and other search engines, Large agencies win every time, and the money to do PPC advertising eventually comes from your reduced commissions and further lowering prices to compete with other agencies.
   
We needed to be more creative in our approach to advertising and think about how we can benefit the creators.  Our smart, learning search and suggestion engine decides the location and placement of the search results (as stated previously,) like shuffling cards, but the deck is not random.  We let the buyers choose the arrangement of images using our internal ranking score.   

The goal is to help creators make better decisions, by giving them data, on how their images perform and providing insights into the industry, etc.. This is something agencies do not do.  I want to do this in real-time.

We try to push the keyword spammers in both organic and paid search down in the search results. We grade images per image buyers needs, and per keyword, image, collection and also added a visual component to check for relevancy of keywords and content.  It is my ultimate goal not to need keywords attached to images for image search but that is a couple of years off.   

NOTE to photographers: If you are searching for a random word that you "loaded" in your images keywords, (not in the dictionary or on the map) to specifically find your images, it is probably not searchable on PicturEngine, or if it is, it won't be for long.  Also if you are searching for your name, we also try to remove that from the searchable content (unless you are selling direct.)  We do this to prevent price comparisons. 

It has taken me and my team years designing and developing our search to listen to the cues of image buyers.  It is not perfect and is getting better with every search.  With that said, we have to be very selective in who we let train our engine, and we had to put qualifiers in place.  We discovered some undesirable results (the hard way) when everyone (not just image buyers) was allowed to teach our engine. 
Example: Images across many subjects began to rise prematurely due to clicking on images to see details or larger images.  i.e. we saw pictures of "waterfall" at the top of our search results start to have more girls with bikinis in them, not the desired result we were seeking.  We needed to adjust our ranking techniques; this is all a learning process.  We are solving the issues in search and ranking as they come, deciding who can and cannot train our engine. 

Yes, some undesirable results will show for a few searches, we try to catch them and adjust, in time they will be filtered and ranked lower where other more relevant images will rise to the top.

The photographer platform is a full-service platform (created and used by my agencies for many years) it includes everything you need for licensing photos (and more) plus the advertising component.
Photographers asked for it, so we provided it as an option.  The only thing the platform does not give is a front facing website to represent your shooting work.  Other platforms like PhotoShelter and PhotoDeck do an excellent job of making websites.  I did not want to focus on website customization and instead focus on a better image search. 

More on that here:
http://picturengine.desk.com/customer/en/portal/topics/187968-photographer-platform/articles
I am trying to keep my post here shorter.

Yes, your images are probably already in our organic search results if they are listed with an agency.  Using our paid Advertising or Platform replaces your images within our search results, and sends the buyer straight to your site or selling page where you make/keep more money. 

With enough support, we can further enable creators to continue to create. 
Tell me, are you happy with the current state of the industry?  If not, what can we change to make it better?


 

Related Topics

  Subject / Started by Replies Last post
39 Replies
19269 Views
Last post March 30, 2009, 00:08
by sgcallaway1994
25 Replies
14392 Views
Last post April 01, 2009, 06:15
by CofkoCof
HEY... It's Working

Started by WarrenPrice Cutcaster

19 Replies
11331 Views
Last post February 23, 2011, 17:15
by stockastic
1 Replies
5194 Views
Last post January 29, 2013, 13:11
by 08stock08
1 Replies
6460 Views
Last post February 09, 2014, 06:38
by Beppe Grillo

Sponsors

Mega Bundle of 5,900+ Professional Lightroom Presets

Microstock Poll Results

Sponsors