Programatic Advertising - for Steemit.com

작년

What is programmatic advertising ?

When companies like Yahoo! started off the concept of advertising was very new. Someone somewhere compared the "page views" to front page, inner pages and other areas in the news papers and niche magazines the age of internet advertising was started. I am not sure where and when this started so I will skip that part. Essentially, we had images or GIFs shown in reserved spaces of numerous sizes. Deals were arbitrarily inked based on the skills of the advertising sales people. This era can be seen as the dawn of "banner" advertisement.

Soon companies started giving out specially targeted advertisements and charge based on regions, time of the day etc. Various standardization efforts like IAB has created definitions, fair pricing strategy and sizes like 300x250 for the ad-banners.

Companies like Google soon started targeting based on not just "geo-fences" but they started grouping users according their interest, behaviors and income groups. As the internet usage exploded, the available "real estate" for showing advertisement increased.

From lets say, mid 2000s we have a scenario where number of "advertisement buyers" aka ad-agencies are competing for premium targeted space. At any point in time there were way too many advertises (Buyers), Sellers (publishers) & Middle-men.

The number of parameters to choose from increased. Now before showing every advertisement, those who serve the ads were comparing against

  • geo-location
  • time
  • income group
  • age
  • sex
  • sexual orientation
  • buying behavior
  • buying behavior of friends and family
  • events in life

And much more decision parameters. Numerous services popped up offering speed, efficiency, anti-fraud technologies. The result was the simple dynamic banner ads replaced by persolized ads matched against numerous rules and selected from a group of competing publishers.

The rule engines, data processing engine for collection and aggregation, matching and finally delivery where all combined and addressed as Programmatic Advertising.

Real Time Bidding (RTB)

An extensive list of technologies in and their definitions can be found here : https://cloud.google.com/solutions/infrastructure-options-for-building-advertising-platforms#terminology

.
As the complexity of serving ads as well as the competition increased, fast, real time decision making and ad-serving using AI, CDN and massive data sets decided the winners. Google, Appnexus and The Rubicon Project emerged as the winners. Needless to say Facebook and Atlas Platform has carved its on niche among all this.

In my view the entire ad-serving still remains looking at the server logs, comparing against existing data and making decisions in near-real time & is defined as Real Time Bidding.

Header Bidding

--
The RTB infrastructure has some of most complicated looking architectures one will come across. They store enormous amounts of data, make decisions at the zillionth of a second, but still the experience was poor and slow.


(Well, just to scare!)

This is when a simple mechanism to contact the backend servers before the content of the page is loaded was devised. This simple "trick" ensured that before the content is even transferred to the users browsers the decision on which advertisement to be shown was taken.

The Steemit Potential

The portal Steemit.com provides a unique audience which can be categorized as early adopters, innovative at the least. Obviously further classifications can be made. Though I don't have the exact numbers, looking at the pubicly available data, steemit.com is a very lucrative medium for the advertisers.

How to do it ?

On a high level, RTB with Header bidding enabled with pass-back tags is the way to go. RTB and Header Bidding are no-brainers. Making use of the pass-back tags ensures that we have a high efficiency interms of how many percentage of the page views has relevant ads. Needless to say, a property like Steemit deserves very high, say, 70 - 80% fill rate. In simple terms, 700 to 800 pages out of 1000 pages will have advertisements.

Where to start ?

My personal favorite is TheGuradian.com. They have an excellent UX & ads are seldom shown and are mostly relevant.

(1. Top banner, 2. The header bidding is activated)

A summary of the ad providers are shown below. I think Google is shown as the last one due to the low rates. The comment also talks about not using a bidding wrapper. My best guess is they are handling the header bidding manually and choosing the networks based on per-determined priorities as opposed to 100% programmatic. I feel this is the BEST approach if we have someone constantly monitoring and tweaking the ad-delivery performance.

Quick Analysis of Steemit.com

2019-01-28_01-28-29.jpg
Currently steemit.com is configured to show a top level horizontal banner and a Below The Line (BTL) vertical advertisement. Its important to move the vertical banner to the top of the page (ATL) to ensure better rates. Considering the fact that we can't show 300x250 ads easily, this is a must IMHO.

Further, we need to enable Header Bidding right now !

Without any specific setting I am seeing the top banner not filled. One strategy is to keep the rates low for the top banner in the RTB or create a private market place in the SSP platform and "lure" advertisers in. Once the AI networks are aware of the advertising space with specific targeting and demographic, we can increase the prices and also go for in house ad selling.

Paid Content

There are portals like Earn.com doing peculiar paid content and advertisements triggering engagement. Such a model is IMHO the best model for Steemit.com as we have vibrant community.

A common model of paid content is the following from The Guradian:

Selling of Merchandise:

Selling of items/SKUs/utilities interesting to the target audience using https://www.stackcommerce.com/ is a method used by Slashdot.org Personally I have found this as the least annoying advertising means.

Conclusion

Targeted and useful advertisements can bring in revenue to a portal with such unique, organic content like steemit.com I hope this article gives pointers to anyone interested.

Actionable list

Here is list of items that can act like a quick reference. I will update with any inputs in the comments too.

  1. Enable Header Bidding
  2. Get approval from multiple ad-networks : Appnexus, Rubicon project & others
  3. Optimize for maximizing the fill rate
  4. Use header bidding script in a "water fall model" as opposed to a wrapper (@techslut suggested "waterfall")
  5. Google Adsense if using should be the last pass-back tag in the waterfall to maximize fill rate
  6. Optimize the page lay out for more Above the Line ads
  7. Change the layout for ads in the promoted content: More like Native Ads (@techslut)
  8. Integrate paid content model
  9. Integrate with StackCommerce or similar services
  10. Explore models like Earn.com
Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
STEEMKR.COM IS SPONSORED BY
ADVERTISEMENT
Sort Order:  trending

Yes! Yes yes yes! I am happy someone who is not me is taking the time to explain how advertising on Steemit can be beneficial to everyone and very profitable for Steemit. If they go as far as create direct relationships with relevant brands and advertisers to put in their waterfall, and add more ad networks that are more relevant to the target audience - AdSense will catch the leftover and all that Steemit traffic can be monetized while the audience gets relevant content.

That said, I am still in favor of a steemit-based own content in-feed promotion system as a source of revenue. How cool would it be if promoted content in-feed was tagged as such and the tokens "burned" like the "Promoted" is supposed to work? Burning token raises value, and people are already using bidbots and votebots for promotional purposes. Bidbots won't die because other frontends and "trending" still exist, but in-feed post promotion is an untapped revenue source, IMO.

·

Thank you for the response :-) ... Someone is listening ...

promoted content in-feed

I think you just mentioned "Native ads" with a twist. I agree that it sounds cool and will have positive impact.

relationships with relevant brands and advertisers to put in their waterfall

This is also needed. I just went for all "programmatic" as there is no dedicated resource at the publisher's end to make this happen. On the waterfall, I agree to this too and with with header bidding too, we can use water fall. The trick is to keep the best performing advertising network with which we have good relationship above everything including Google Analytics (GA has demographic and data sharing which can be enabled) to get the best possible rates. Google adsense or ADX even if we put them at the end are normally very fast because they are mostly deployed at the "edge".

Tokens "burned" like the "Promoted" is supposed to work?

I didn't get this - you mean to say if the paid content in the promoted or other feeds if gets voted their resulting rewards can be burned ?

·
·

I didn't get this - you mean to say if the paid content in the promoted or other feeds if gets voted their resulting rewards can be burned ?

Not the rewards, just the money paid to be promoted. The promoted posts (fine fine fine native ads) then "compete" for views in the relevant tag feed or general trending feed. I've offered this before many times: simply take the existing "Promoted" mechanics and page, and spread it throughout the feed like Reddit does. I want to see SBDs burning here!!!

·
·
·

Got it. Thanks for the explanation

Excellent article Bobby. I think suggestions from this article should be taken as one of the enhancement scope for the front ends we have for Steem.

Advertisements is huge subject that is gaining more focus for the past decade. Data science is also collaborating with the advertising concept to make it one of the hot courses to study in top universities.

Posted using Partiko Android

·

Thank you :-)

Yes adverting has become nothing but one of the most important applications of data science. Whether we like it or not, some of the best minds on earth are working in the "Advertising business". We all know the big advertising companies - Google, FB :-)

This post has been rewarded with 100% upvote from @indiaunited-bot community account. We are happy to have you as one of the valuable member of the community.

If you would like to delegate to @IndiaUnited you can do so by clicking on the following links: 5SP, 10SP, 15SP, 20SP 25SP, 50SP, 100SP, 250SP. Be sure to leave at least 50SP undelegated on your account.

Please contribute to the community by upvoting this comment and posts made by @indiaunited.

Imagine a steem interface (better version of steemit) where content creators could decide if they want to include ads like this in their posts. Some small part of the revenue could even go to steem inc.

So much potential but @ned is just completely not interested in making steemit a viable social platform that generates income. No idea why:(

Also great post, thank you:) !tipuvote 5

·

Appreciate the nice words @cardboard :-)

Some small part of the revenue could even go to steem inc.

The challenge with allowing ads from end-users is the potential for abuse. Even with all the anti fraud and scanning measures we end up having ads from different categories. Imagine ending showing ads related to Tobaco to younger audience or violate the geographical locations. This can get very tricky.

One of the funniest and toughest scenario that I have found myself is "defining nakedness". We manage ads for a news paper owned by a certain religious group. The news paper often quite popular and unless they are reporting matters close to their religion, they are very good. They have strict policy against tobaco, liquor and also no dating advertisements. Occasionally an advertisement pops up which may be for something like a matrimonial site. Some takes a screen shot and shares with the management. Someone from our end gets a call. Once the representative from our end "declined" that the particular ad contains no nudity or nakedness. This got escalated and we found ourselves in the business of defining "nakedness" and "level of nakedness" allowed in the religious context - We finally had to apologize, give them discounts and steer away.

Apart from the uncertainty like this, allowing ads from the users also will have challenges like the one which Myspace faced - extreme customization was allowed resulting in challenges.

But the idea of giving back to the "authors" as an ongoing revenue is something that has immense potential. With the current RTB servers it is possible thogh the return will be low as the payouts are something like max 10 USD per 1000 impressions.

A solution like what is being offered by BAT (basic attention token) can be explored. I think @themarkymark had some ideas around BAT. I am interested in hearing thoughts from anyone who has ideas about BAT + STEEM :-)

·
·

Thanks for the answer :) Check out @dclick app - it's something like a first version of users generated ads on steemit.

I think it would be cool to create something like a ads market - where you can both create an ad and pick which one you want in your post. Not my idea though!

Every single word in here must be taken seriously. You are totally right man. Great analytics!

  ·  작년

This post is supported by $2.35 @tipU upvote funded by @cardboard :)
@tipU voting service guide | For investors.

This post has been included in the latest edition of SoS Daily News - a digest of all you need to know about the State of Steem.



Interesting proposals. We're definitely excited to see @steemit start optimizing their monetization models if that means more development for the Steem Community as a whole.