[am1m]

Originally Posted by anjan_c2007

"A radio station still run by humans."

Great tag-line, great differentiator! (And thanks, will probably use some version of that somewhere!)

To add to the discussion however, the issue is not something that boils down to just human-run/generated vs. automation. As far as content goes, there are a lot of aspects that benefit from automation and actually enhance the user/reader experience. But for that to work, the basic content you work with has to be good.

The problem is when a site's basic content is basically crappy. If you're working with bad, poorly researched content, no amount of polishing or slick production can disguise that for very long. That brings us to a fundamental issue that has plagued newspapers for a while and is now becoming apparent in websites- no one wants to pay for content. Or rather- when bad content is available for free, most people would rather consume that and very, very few will pay for quality content. (These days it's gone a step lower- most people rely on whatsapp forwards to get news and won't bother verifying if what they are reading is really true.) That in turn leads newspapers, magazines and websites to pay content creators and photographers lower. Why pay a qualified person to create content when you can instead either: 1. Pay 10-20 unqualified creators to generate volumes of content, however crappy or 2. Pay a celebrity influencer who will get clicks and likes.

In the news world, advertising subsidizes news coverage. But the issue of getting subscribers to pay more for a better newspaper is proving difficult to solve. So till a seamless model is found, we'll have to suffer 'news' like "xyz actress spotted at abc place wearing 123 outfit" or "xyz cricketer spotted with his family at 123 destination" (all mostly paid-for PR placements btw, especially if the star is a b-lister or new to the field). And we'll have to suffer through auto-generated comparison lists online.

[MaheshY1]

As someone who does this, here's my justification:

1. the value for the reader in auto-generated comparison articles lies in the tables, not the sentences. The sentences are for google. Tables are easily glanceable.
2. Google, its assistant and the home don't look at tables and provide the answers. They read, from our auto-generated sentences. Because of this, we can gradually get to a point where you won't have to pick up a device, open the browser and head to the manufacturer's website for reliable info.
- You can just ask which variant of a particular car gets 6 airbags.
- Or what's the most affordable car with cruise control.
- Or mention 2 cars and ask which one has better ground clearance/boot space/wheelbase, etc.
Neither Google nor the content creators are there yet to answer all of these questions, but it's getting there.

The buyer can read through a long ownership report or official review on TBHP. But we are a small number of enthusiasts. We can't expect an average Joe to do that kind of lengthy research.

3. There are obvious cost advantages for teams that use automation. While the content can be created instantly once the code is written, the code writing and testing itself doesn't happen overnight. But in the long run, it comes very close to delivering 80% results with 20% effort.

4. Simplicity - Just because we can convey a point with a fancy sentence that only a human can write, doesn't mean we always should. Just like Google, I love simplicity too. Because I haven't ever heard anyone complaining about something being too easy to understand.

E.g. 1: "Car A has 200mm of ground clearance, which is 10mm better than Car B."
vs
E.g. 2: "Car A's 200mm ground clearance is the benchmark for the sub-4m SUV segment. Car B's 190mm clearance is almost there but not enough to convince the buyers who consider 200mm as the baseline to call something an SUV."

The 1st example assumes that the reader knows the importance of GC and is just interested in comparing the GC of 2 cars.
The 2nd example uses 170% more words to convey pretty much the same info. It takes longer to read it. it assumes that the reader is not aware of what minimum GC he should expect. The additional info is irrelevant for most of the readers, who just want to compare. A human will have to create content that empowers the buyers to make informed decisions. Then, this automated content can help the buyer in filling the information gaps before zeroing in on a vehicle.

[sinhead]

Originally Posted by Rehaan

• The problem is that Google laps these pages up, since they've been specifically designed to appease their algorithms.

Will using some other search engine be better then?

[MaheshY1]

Originally Posted by sinhead

Will using some other search engine be better then?

I tried Ducking it, and Zig was towards the top. The automated content wasn't the thing that bugged me here.



The problems I notice here are:
1. third word P capital
2. L capital in lakh
3. Figo FE is 23.5kmpl, mentioned 24.4kmpl



4. Consistency with units. Zig uses PS for power but here it's BHP for power, which will confuse buyers.
5. lack of depth in content.

In technical specifications, Ford Figo (Top Model) is powered by 1499 cc engine , while Ford Freestyle (Top Model) is powered by 1499 cc engine

For this, they needed another 'IF' condition when the displacement is the same. Then another 'IF' when the power and torque are the same. Then another 'IF' when the peak torque is not 'at' an RPM but is in a range. E.g.:

Both, the Figo and Freestyle diesel use 1499cc turbo diesel engines. They also have identical power and torque output of 100PS at 3750rpm and 215Nm between 1750 and 2500rpm, respectively.

This works even if we don't want to get into whether these car have only the same displacement, power and torque or they're literally the same engine.

6. They chose the diesel models and the top variant by default without asking me which variant I want. If I wanted petrol, then I'm very likely to ignore this page altogether.
7. inconsistency in engine type: one is TDCI of unknown displacement, other is a 1.5L diesel engine.
Also, see the engine of Venue turbo in the following image: it's mentioned as 1.2 MPI with a displacement of 999cc.



For most media brands, the story will be similar.

So, the automated-ness is less of a problem here if the tables are done right and the data is accurate. Among the brands doing this, clearly, non-car people thought it up and are doing this and maintaining the model pages. That's why the discrepancies.

An even bigger problem than the automated content is the non-automated, biased, human-written content. With favors, media gifts, free tours, review cars, it's very easy to buy out an opinion and paint it green. I'd rather take an automated comparison that tells me that the XUV300 has the same GC as the Tata Nano over one that devotes a paragraph to its largest-in-segment wheel size -- without letting people know the impact of rotating mass on fuel efficiency, braking and acceleration.

Disclaimer: I have no hard feelings for Figo, Venue, XUV or Zig. It just happened to pop on top, so, I thought of taking a look and seeing where it went wrong to have aroused such strong negative emotions about automated content.

[Nitish.arnold]

Originally Posted by shrinz.vivek

Many such instances makes me wonder, is the internet filled with auto generated articles? If yes, how authentic is it? Like many other skills will we lose our essay writing or comprehension skills sooner than expected? What will be the impact of this on future generations who grow up reading such material?

Hey there, These are widely used practices, Source- Digital Growth Hacker this side

One example, with the technicalities explained :
Sometime back, I used to freelance as a consultant for a leading mobile/electronic review platform in the APAC region,
The tech team used just one templatised content piece of ~1000 words or so, with placeholders for the model, specifications, size and more, which were dynamically called from the database.

one line example of that below:-
<The [mobile name] is available in [available colors] finishes, in [ram capacity] RAM and [storage capacity] storage configuration priced at Rs.[price] & affiliate link.>

The vs articles were more simple to generate,
the team generated ~5 crore articles and the SEO (organic) traffic shot up like anything.

------------------------------------------------------------------------

However, Google's algorithm is getting smarter day by day, giving a lot of weightage to metrics like user signals (engagement, Time on site, Pages per session, bounce rate etc.),
& these tactics won't last much.

User engagement is very important for any platform to build credibility,
plus their is a unique voice of every author with which the audience relates a lot (voice = style of narrating).

So, to answer your query in short, don't worry so much about these articles.

[Rehaan]

Originally Posted by sinhead

Will using some other search engine be better then?

Good question, and I guess my answer is that it will be a gamble.

I'd imagine anyone trying to optimize their content for search engine rankings would have Google as their #1 search engine by far.

Therefore you could say the manipulation isn't geared towards other search engines as much.

However, that doesn't mean they aren't going to be affected by it...