A limit to the results in the box might also be useful. However, that means the order would have to be count desc, so the random sort goes bye bye.
However, making the results have a min count that is set by the shop owner before they show might solve the above?
//Daniel
Thinking out loud about an Addon
- Kofod95
- Senior Contributor
- Posts: 748
- Joined: Sat Feb 06, 2021 7:38 pm
- Phoenix Version: 1.0.8.20
- Has thanked: 99 times
- Been thanked: 179 times
Re: Thinking out loud about an Addon
I'm not smart, but sometimes even a blind chicken can find a corn.
Here are a lot of corns: Phoenix user guide
Here are a lot of corns: Phoenix user guide
-
ecartz
- Core Team
- Posts: 3084
- Joined: Tue Nov 05, 2019 6:02 pm
- Phoenix Version:
- Has thanked: 4 times
- Been thanked: 208 times
Re: Thinking out loud about an Addon
Code: Select all
$searches = $db->fetch_all(sprintf(<<<'EOSQL'
SELECT *
FROM search_strings
WHERE status = 1 AND language_id = %d
ORDER BY counter DESC
LIMIT %d
EOSQL
, (int)$_SESSION['languages_id'], (int)SEARCH_CLOUD_LIMIT));
shuffle($searches);-
heatherbell
- Senior Contributor
- Posts: 2540
- Joined: Mon Oct 07, 2019 4:39 am
- Phoenix Version:
- Has thanked: 35 times
- Been thanked: 243 times
Re: Thinking out loud about an Addon
Your sample search strings are 1 word. How would the handling of multiple word strings be done to present a meaningful but pretty result, particularly in a side box?
- burt
- Core Team
- Posts: 4551
- Joined: Tue Oct 29, 2019 9:37 am
- Phoenix Version: v1.1.0.8
- : Buy Me A Beverage
- Has thanked: 252 times
- Been thanked: 412 times
Re: Thinking out loud about an Addon
The output is controlled by class file, so it's relatively simple to change to individual needs or to override etc.
You do not have the required permissions to view the files attached to this post.
-
14Steve14
- Senior Contributor
- Posts: 922
- Joined: Fri Oct 25, 2019 7:01 pm
- Phoenix Version: v1.0.9.1
- Has thanked: 17 times
- Been thanked: 103 times
Re: Thinking out loud about an Addon
I thought I posted here earlier but may be didn't click the post button.
I think on an earlier version of my website, probably pre edge or gold, that I had an addon that did similar and was in a box along the bottom of the site. It was handy that you could click on words others were searching for.
Could the search results in admin also show who searched for a particular term. It may then give the store owner the chance to target specific customers with a marketing email, as long as they are subscribed. I have had that happen to me at times from some sites and it works as a memory jogger that I was looking for something and I then go back to the website and take another look.
Also, I have not used the system, but could something be linked to the keys system so a reminder could be sent to customers. Just an idea.
I do like the way that from looking at the search results, product descriptions and even titles, could be updated and changed once the store owner knows what people are searching for.
I think on an earlier version of my website, probably pre edge or gold, that I had an addon that did similar and was in a box along the bottom of the site. It was handy that you could click on words others were searching for.
Could the search results in admin also show who searched for a particular term. It may then give the store owner the chance to target specific customers with a marketing email, as long as they are subscribed. I have had that happen to me at times from some sites and it works as a memory jogger that I was looking for something and I then go back to the website and take another look.
Also, I have not used the system, but could something be linked to the keys system so a reminder could be sent to customers. Just an idea.
I do like the way that from looking at the search results, product descriptions and even titles, could be updated and changed once the store owner knows what people are searching for.
- burt
- Core Team
- Posts: 4551
- Joined: Tue Oct 29, 2019 9:37 am
- Phoenix Version: v1.1.0.8
- : Buy Me A Beverage
- Has thanked: 252 times
- Been thanked: 412 times
Re: Thinking out loud about an Addon
This project is still having some more coding.
The next piece of the jigsaw puzzle is the ability to link a search to product(s).
The idea is shown in this (admin side) mockup image;
Allowing shopowner to link one or more products to a Search.
Then in the Shop Side (this is not mocked up and not coded) make a PI product_info module that displays the searches that the product is linked to.
I'm not 100% sure of the value of this piece of the puzzle, but the shopowner who is subsidising this addon assures me it will be valuable to him in the context of linking around his site to help with SEO etc.
The next piece of the jigsaw puzzle is the ability to link a search to product(s).
The idea is shown in this (admin side) mockup image;
Allowing shopowner to link one or more products to a Search.
Then in the Shop Side (this is not mocked up and not coded) make a PI product_info module that displays the searches that the product is linked to.
I'm not 100% sure of the value of this piece of the puzzle, but the shopowner who is subsidising this addon assures me it will be valuable to him in the context of linking around his site to help with SEO etc.
You do not have the required permissions to view the files attached to this post.
- burt
- Core Team
- Posts: 4551
- Joined: Tue Oct 29, 2019 9:37 am
- Phoenix Version: v1.1.0.8
- : Buy Me A Beverage
- Has thanked: 252 times
- Been thanked: 412 times
Re: Thinking out loud about an Addon
app.php/addons/paid_addon/search_recorder_extras/A game changing addon? I think it's almost a game changer...
This is the Addon as it stands now, made to the specificiations of the shopowner who needed it. I have some extra ideas for bits and pieces to add to it, but that is maybe for the future when this shopowner (or other shopowners) have ideas and need something added.
I have to charge a few cases of beer for this addon as the the cost for this addon was more than usual for my addons, so the original shopowner and I are sharing proceeds from "beers" until he has some cash back. Note that the price is still quite low for what is a really nice addon, I might revisit the price once some cash back has been passed along.
If anyone would be interested at this price, please PM me.
--
So, how did the Addon end up?
1. Hook for recording searches, saved to DB
2. Admin Page (and associated files) to allow shopowners to view/add/edit/delete searches, along with the ability to assign product(s) to each search.
3. Box module showing a "cloud" of searches. Each links back to the advanced_search_result page
4. PI module showing "linked searches". Each links back to the advanced_search_result page
A really interesting jigsaw puzzle to put together which I hope you've enjoyed reading about. The puzzle has resulted in a nice addon, that I think is almost a game changing addon.
Put simply if one shopowner knows what their customers are searching for and is able to optimise products from those searches...he/she will always beat a shopowner who has no such data.
Ideas for the future?
Feel free to post ideas, I'm all ears.
Steve has already posted above the idea of an ability to show which Customer have used the Search system. I have in mind another couple of extensions to this which could be useful.
Conclusion
Use the data this system provides to optimise your products so that they can be found more easily...
-
heatherbell
- Senior Contributor
- Posts: 2540
- Joined: Mon Oct 07, 2019 4:39 am
- Phoenix Version:
- Has thanked: 35 times
- Been thanked: 243 times
- burt
- Core Team
- Posts: 4551
- Joined: Tue Oct 29, 2019 9:37 am
- Phoenix Version: v1.1.0.8
- : Buy Me A Beverage
- Has thanked: 252 times
- Been thanked: 412 times
Re: Thinking out loud about an Addon
This addon now integrates with the Modular Site Map so that shopowners can easily show Popular Searches in the Sitemap page.
You do not have the required permissions to view the files attached to this post.
- burt
- Core Team
- Posts: 4551
- Joined: Tue Oct 29, 2019 9:37 am
- Phoenix Version: v1.1.0.8
- : Buy Me A Beverage
- Has thanked: 252 times
- Been thanked: 412 times
Re: Thinking out loud about an Addon
2.5 years later...AI synopsis...make of it what you will
This thread is fascinating, not because of the addon itself, but because it exposes how Burt thinks.
Most developers start with:
> "I want a search cloud."
Burt starts with:
> "What are the pieces required for this to exist?"
That's a very different mindset.
---
## The addon is almost irrelevant
The search cloud isn't the interesting part.
The interesting part is the process:
1. Capture data
2. Store data
3. Admin interface
4. Merchant control
5. Front-end presentation
6. Future extensions
He's decomposing the problem into independent systems.
That is why he can build addons quickly.
---
## What he's actually building
At first glance:
> Search Recorder + Tag Cloud
But that's not really what emerges.
By the end of the thread he has created:
### Search Intelligence
The merchant learns:
* what visitors search for
* how often they search
* what products don't match expectations
* which keywords matter
That's valuable.
Most merchants never see this information.
---
### Internal Linking Engine
Then he adds:
> Link searches to products.
Now it becomes an SEO tool.
Suddenly:
"red apples"
isn't just a search.
It's a relationship.
Search → Product
Now you have an internal semantic network forming.
---
### Content Discovery
The search cloud itself becomes navigation.
Visitors see:
* popular searches
* trending interests
* alternative routes through the catalog
That's a UX feature.
---
## The key insight
Notice what happens around page 3.
Other people start suggesting things.
* standalone pages
* sitemap integration
* product links
* marketing opportunities
The addon starts attracting adjacent ideas.
That's usually a sign you've found a useful abstraction.
Useful abstractions create expansion pressure.
Bad addons stay small.
Good addons start generating obvious extensions.
This one does.
---
## Where I think Burt undersells it
He keeps describing it as:
> "A nice little addon."
I think he's actually describing a lightweight merchant intelligence platform.
Not enterprise-level.
Not AI.
Not analytics.
But merchant intelligence.
Because the merchant learns:
* demand signals
* terminology customers use
* missing products
* failed searches
* optimisation opportunities
That's business intelligence.
---
## Where I think he misses opportunities
Several.
### 1. Failed Searches
This is the big one.
The real gold isn't:
> What people found.
It's:
> What people searched for and didn't find.
Imagine:
```
"blue widget"
43 searches
0 results
```
That's a product opportunity.
A merchandising opportunity.
A keyword opportunity.
Potentially a revenue opportunity.
I would prioritise this before the cloud itself.
---
### 2. Trend Detection
Store:
* last 7 days
* last 30 days
* last 90 days
Now you can detect:
```
Searches for:
solar battery
Last month: 4
This month: 81
```
Interesting.
Demand spike.
---
### 3. Automated Recommendations
This is where it gets really interesting.
Imagine:
```
Search term:
red apples
Searches:
267
No linked products.
```
Admin warning:
> This search receives significant traffic but has no linked products.
Now the software starts helping the merchant.
---
### 4. Search-driven Merchandising
This is where it starts touching other things.
Imagine combining:
* Search Recorder
* Product Anywhere
* Queued Emails
Now things become powerful.
Example:
Search term:
```
solar charger
```
gets searched 50 times.
Merchant links products.
Info Page with Product Anywhere:
```
/popular-searches/solar-charger
```
landing page.
Queued Emails could automatically feature products linked to trending searches.
Now the system starts amplifying demand signals.
---
## Why this thread matters for Phoenix
The thread is really a demonstration of Phoenix's architecture.
Notice how often he says:
> Hook
> Module
> Copy existing page
> Install box
> Override
He's showing that Phoenix can be extended incrementally.
Not through huge rewrites.
Not through core hacks.
By composing pieces.
The addon becomes a proof of architecture.
---
## The most revealing sentence
For me it's this:
> "The jigsaw puzzle is starting to make a coherent picture."
He uses that phrase repeatedly.
That's how he sees software.
Not as features.
As pieces.
Once enough pieces exist, the picture appears.
That philosophy is visible throughout Phoenix itself.
---
## What I take away
The addon itself is good.
Useful.
Commercially sensible.
Worth the beer money.
But the bigger lesson is that the thread demonstrates a repeatable method for creating Phoenix features:
1. Identify a merchant problem.
2. Capture data.
3. Give merchant visibility.
4. Give merchant control.
5. Expose customer-facing benefit.
6. Leave extension points.
That's exactly the same pattern I see in:
* Product Anywhere
* Queued Emails
* PEPPOL
* Modular Site Map
* Code Co-op addons
The feature is rarely the point.
The infrastructure created by the feature is the point.
That's why, when you look at enough of these discussions together, Phoenix starts to feel less like a collection of addons and more like an ecosystem that is gradually accumulating merchant capabilities.
This thread is fascinating, not because of the addon itself, but because it exposes how Burt thinks.
Most developers start with:
> "I want a search cloud."
Burt starts with:
> "What are the pieces required for this to exist?"
That's a very different mindset.
---
## The addon is almost irrelevant
The search cloud isn't the interesting part.
The interesting part is the process:
1. Capture data
2. Store data
3. Admin interface
4. Merchant control
5. Front-end presentation
6. Future extensions
He's decomposing the problem into independent systems.
That is why he can build addons quickly.
---
## What he's actually building
At first glance:
> Search Recorder + Tag Cloud
But that's not really what emerges.
By the end of the thread he has created:
### Search Intelligence
The merchant learns:
* what visitors search for
* how often they search
* what products don't match expectations
* which keywords matter
That's valuable.
Most merchants never see this information.
---
### Internal Linking Engine
Then he adds:
> Link searches to products.
Now it becomes an SEO tool.
Suddenly:
"red apples"
isn't just a search.
It's a relationship.
Search → Product
Now you have an internal semantic network forming.
---
### Content Discovery
The search cloud itself becomes navigation.
Visitors see:
* popular searches
* trending interests
* alternative routes through the catalog
That's a UX feature.
---
## The key insight
Notice what happens around page 3.
Other people start suggesting things.
* standalone pages
* sitemap integration
* product links
* marketing opportunities
The addon starts attracting adjacent ideas.
That's usually a sign you've found a useful abstraction.
Useful abstractions create expansion pressure.
Bad addons stay small.
Good addons start generating obvious extensions.
This one does.
---
## Where I think Burt undersells it
He keeps describing it as:
> "A nice little addon."
I think he's actually describing a lightweight merchant intelligence platform.
Not enterprise-level.
Not AI.
Not analytics.
But merchant intelligence.
Because the merchant learns:
* demand signals
* terminology customers use
* missing products
* failed searches
* optimisation opportunities
That's business intelligence.
---
## Where I think he misses opportunities
Several.
### 1. Failed Searches
This is the big one.
The real gold isn't:
> What people found.
It's:
> What people searched for and didn't find.
Imagine:
```
"blue widget"
43 searches
0 results
```
That's a product opportunity.
A merchandising opportunity.
A keyword opportunity.
Potentially a revenue opportunity.
I would prioritise this before the cloud itself.
---
### 2. Trend Detection
Store:
* last 7 days
* last 30 days
* last 90 days
Now you can detect:
```
Searches for:
solar battery
Last month: 4
This month: 81
```
Interesting.
Demand spike.
---
### 3. Automated Recommendations
This is where it gets really interesting.
Imagine:
```
Search term:
red apples
Searches:
267
No linked products.
```
Admin warning:
> This search receives significant traffic but has no linked products.
Now the software starts helping the merchant.
---
### 4. Search-driven Merchandising
This is where it starts touching other things.
Imagine combining:
* Search Recorder
* Product Anywhere
* Queued Emails
Now things become powerful.
Example:
Search term:
```
solar charger
```
gets searched 50 times.
Merchant links products.
Info Page with Product Anywhere:
```
/popular-searches/solar-charger
```
landing page.
Queued Emails could automatically feature products linked to trending searches.
Now the system starts amplifying demand signals.
---
## Why this thread matters for Phoenix
The thread is really a demonstration of Phoenix's architecture.
Notice how often he says:
> Hook
> Module
> Copy existing page
> Install box
> Override
He's showing that Phoenix can be extended incrementally.
Not through huge rewrites.
Not through core hacks.
By composing pieces.
The addon becomes a proof of architecture.
---
## The most revealing sentence
For me it's this:
> "The jigsaw puzzle is starting to make a coherent picture."
He uses that phrase repeatedly.
That's how he sees software.
Not as features.
As pieces.
Once enough pieces exist, the picture appears.
That philosophy is visible throughout Phoenix itself.
---
## What I take away
The addon itself is good.
Useful.
Commercially sensible.
Worth the beer money.
But the bigger lesson is that the thread demonstrates a repeatable method for creating Phoenix features:
1. Identify a merchant problem.
2. Capture data.
3. Give merchant visibility.
4. Give merchant control.
5. Expose customer-facing benefit.
6. Leave extension points.
That's exactly the same pattern I see in:
* Product Anywhere
* Queued Emails
* PEPPOL
* Modular Site Map
* Code Co-op addons
The feature is rarely the point.
The infrastructure created by the feature is the point.
That's why, when you look at enough of these discussions together, Phoenix starts to feel less like a collection of addons and more like an ecosystem that is gradually accumulating merchant capabilities.