How to create meta descriptions for category pages in Word Press

I have been spending a lot of time going SEO for my company site Mingo Smart Factory. When revamping the site architecture; I created a category hierarchy and category pages. 

Using Yoast to Set Meta Description of a Category Page

By default, Yoast uses the post excerpt for the meta description. But on a category page there is no post body for the system to create an excerpt from—only the description of a category.

Setting a Custom Meta Description

Below is the Yoast configuration for this category page. You can see that I created a custom description for this page rather than having Yoast generate it.

Using the Category Description as a Default

The other option is configuring Yoast to use the category description as the meta description. When you first install Yoast it is configured to display meta descriptions on category pages. Below is a screen show of the out-of-the-box configuration.

Navigate to Yoast -> Settings -> Categories to see this section. Notice that Excerpt is in the meta description field.

To fix this, delete Excerpt from the box and type %T. Then you see a drop-down of values starting with the letter T, choose Term Description, then save your configuration.

Now you have meta descriptions set for all of your category pages that have descriptions!

July 24th 2021, North Fork of Flathead River

Standup Paddle Boarding the North Fork

Just when you think you know what you are doing the universe kicks you in the teeth to remind you how little you know.

The North Fork of the Flathead River starts in Canada follows the western edge of Glacier National Park. It finally merges with the Middle Fork of the Flathead River right outside of Glacier National Park. Overall this section of the river is known to be pretty chill, no big rapids. But it does have fast moving water, log jams and other dangerous stuff.

We started just north of Polebridge, Montana where there is an easy put in spot by the bridge into Glacier National Park.

We ended at Coal Creek which is a popular take out spot right where Coal Creek feeds into the North Fork River.

  • Lewis & Clark on a Paddle Board

7/4/21 – Swan River, Montana

Standup Paddle Board on the Swan River

Now that we’ve become experts at stand up paddle boarding having spent a total of 5 1/2 hours on the boards on a very chill river I wanted to try something more challenging.

The Swan River what is the place to go. This one was a lot of fun some fast-moving waters not quite what you would call rapids they are called riffles.

Stand Up Paddle Boarding The Whitefish River

Whitefish Montana- 7/27/2021

We bought a couple of Body Glove stand up paddle boards earlier this year. We have a lot of friends in Montana who raft and also stand up paddleboard on all the rivers around us. So we bought some paddleboard’s.

Are the rivers in Northwest Montana can move pretty fast and get pretty crazy so we wanted to start easy and this was the first adventure on the Whitefish River.

Whitefish river moves very slowly you can actually paddle up and down it there are no Rapids there there is no fast-moving water perfect place for our first paddle boarding adventure.

We did the river twice, once starting at Baker Park and paddling up towards the lake. And once starting at Whitefish beach and paddling all the way to Highway 40.

The highway 40 day took over four hours and it was a hot one sunny clear skies 94° but low humidity.

Best part about this trip was we saw a bald eagle flying overhead maybe about 30 feet up in the air. Never seen one that close before and of course we don’t have a picture of it.

The day we went back-and-forth from Baker Park we’re short about an hour and a half relatively slow we paddle just passed second Street and then turned around.

And as usual I managed to fall off the board and totally calm water for no apparent reason.

Wayfarers Park, Flathead Lake State Park

8/8/21 great park on the north east side of Flathead Lake. Beautiful cliffs that look west over the water. The perfect stop to check out how big and beautiful the lake is. I’m late July and early August you can grab fresh picked cherries just south of the park on the side of the road.

View of Flathead Lake from the Cliffs at Wayfarers Park

Underpants and IoT Platforms

A long time ago South Park had an episode call the Underpants Gnomes, it was meant to poke fun at .com companies and the internet bubble. The analogy has been used 1000s of times and I’m going to use it again.

To review the Underpants Gnomes have the following business plan:

  1. Collect Underpants
  2. ??????
  3. Profit

We can modify this to fit IoT Platforms as follows:

  1. Build IoT Platform
  2. ?????
  3. Profit

Field of Dreams

I guess I can use a Field of Dreams reference here too, if you build it they will come. They won’t. I wrote a few times how too many companies are building IoT platforms with no real use case or tangible benefits. (here and here).

They believe if you build an IoT platform people will figure out how to use it and you will make tons of money and dominate your market.

That is what Amazon, Salesforce, Microsoft, Box and every other tech giant did, right? They built a platform which everyone adopted and boom they had a billion dollars in sales. Nope they built a killer application then built the platform on top of that success.

Applications not Platforms

Think about Salesforce, did they start out building force.com, a platform to build any business application in the cloud? No, they built a CRM tool, realized people needed to customize it and can use CRM for many other things besides sales, then built a platform.

That is what the 400 or so IoT platforms need to do. Find a niche where the provide value grow that niche then expand. Then build a platform based on that success.

This is what SensrTrx is doing, we created an application to help manufacturers extract and understand data from their machines to make them run more efficiently. SensrTrx is not a platform, it is an application designed to do just that one thing.

One day it may grow into a platform but for now I’m happy with the progress we have made and the value we provide as an application.

Inflated Expectations

This has less to do with underpants and more to do with marketing. Last week an analyst from Jefferies, an investment banking firm, published a report on IBM. The report said that the promises IBM is making for Watson are falling far short of expectations. IBM commercials show Watson is a complete product for any application, is quick to deploy and has a short time to value. The reality seems to be different.

There is a large services component associated with every project and they don’t always deliver results.

The problem is that IBM has such strong marketing it warps people’s perception of what is possible and real. People assume Watson can write songs with Bob Dylan and cook with the best chefs, but that is not really true. These unrealistic expectations are then applied to any other companies claiming they have AI and machine learning technology in their applications.

The exposure of these technologies to the public is great but over the long term may do more harm than good.

Many of the IoT platforms from the big players in the industry suffer from the same issue of over inflated expectations. Which is just as dangerous, as these projects fail to reach the inflated expectations it will make other companies less likely to invest in the future. Slowing overall adoption.

At some point the platform wars will end and the companies focused on customer value and applications will be left standing. We will see the benefits of all of this technology but we have to wade through the hype to find true success stories.

Better Data Beats Better Algorithms

I really enjoy the SaaStr the podcast and listen every week, the content is usually good but sometimes they hit it out of the park. During an episode a few months ago one of the guest said:

“Better data beats better algorithms.”

Many people debate if more data will be a better algorithm but few talk about how better, cleaner data will beat an algorithm. You can scroll through Google all day on the millions of pages that discuss this more data vs better algorithm.

But when you talk with data scientists and statisticians they will all tell you that you must start with clean data and then run your algorithms.

But if we look across the landscape of companies who are developing predictive analytics products for industrial companies they all collect tons of data and run some magic algorithm on it to give you insights.

Most of these people don’t come from manufacturing. They don’t understand how and why production rates differ from day to day. And are probably going to give you insights that and experienced manufacturing person would already know. So where is the value in these solutions? There are no magic general AI algorithms that can dig through data and give you magic insights.

All of the manufacturing data coming from the factory should be cleaned and enriched as it is stored. This way you have clean data to start from and all the context around what happened with a part was being made, the machine went down or your scrap rates were higher than normal.

If you write machine learning algorithms on this cleaned data you will get much better results. Then if you are blindly trying to find patters and correlations in the data.

This is exactly what my company SensrTrx does.

Sensor Data is Meaningless without Context

Last month I wrote a post for SensrTrx on how sensor data is meaningless without context. It was well received in my network on LinkedIn and twitter. But I find it interesting how many companies are selling just the collection of sensor data.

Products like Power BI for example do this, they want you to throw all of your sensor data into database and graph it. Well thats great but once you start to get beyond a single sensor or a set of related sensors it is worthless. You will start asking yourself the following questions:

  1. What shift was that?
  2. What part were we running?
  3. What else was happing on that machine or production line?

It is possible to get this information from Power BI or Tableau but now you need data from other systems beyond your sensor which means calling IT, more time and more cost. Not to mention that your answer won’t be timely and what ever problem you were looking to solve is no longer important.

Here is a link to my original post: Sensor Data is Meaningless without Context.

Side note: The same can be said for AI algorithms that supposedly correlate a bunch of data points together to help you find patterns you did not no existed. I’m going to tackle this in another post.