Things change rapidly in the WordPress world. The content in this post is more than a year old and may no longer represent best practices.
Thanks to everyone who attended this past Sunday in our new meetup location. Here’s my presentation on Business Intelligence and WordPress, which I’ve uploaded to SlideShare.
Sallie’s Notes on Anca’s Business Intelligence Presentation
Definition of Business Intelligence: The tools and techniques to acquire data and transform into meaningful and useful information.
Some questions that business intelligence is designed to answer:
- Who are my best customers?
- How much money have I made this year?
- What will it take to fulfill an order or provide a service? (COGS and overhead)
- How effective are our marketing campaigns?
- How well are our systems working?
The answers to these questions help you determine which products to keep, whether to add or reduce staff, whether to enter or exit a market, which marketing campaign to undertake
So where does WordPress fit in?
You can use WordPress to manage your organization and centralize forms, content, commerce, appointments/events, services, social conversation, contracts and payments, account management.
Tech Liminal uses a lot of services, and they’re planning to narrow that down. Getting rid of EventBrite, for instance. Every system creates friction in your business. Make a mind map of these for yourself so you can figure out which ones you don’t need.
“We never use our Google Sites ‘Intranet’: it’s a pain.”
Beware of VC-funded startups that are not yet profitable: they may go away and take your data with them. (Lisa LaMagna had this experience with Zirtual.)
Business Intelligence Reports and WordPress
AMR Users plugin will tell you anything at all about your site’s users. It’s very fast to query because it caches tables.
WooCommerce reporting will show you how much you’ve made, and you can export it, but it’s incomplete. There’s always more that you want. There are several reporting extensions for WooCommerce to expand what you get in the basic package.
One of WP’s strengths is the small number of database tables. In terms of what you want to report on, it’s posts, post meta, users, user meta, terms, term meta.
The meta tables help keep the main tables simple. Writing SQL queries goes well when you have a moderate number of users and meta records. But meta tables can really blow up.
If you write plugins, remember that storing data in an array makes it harder to query.
When you get hundreds of thousands of users, your queries start to time out because there is SO MUCH META. For a while you can just increase your hosting, but eventually you have to find another solution.
Creating intermediary tables to make reporting more efficient. It takes about 10 seconds to generate this table for 290,000 users on the Social Media Week site. Intermediary tables do not mess with the original data.
For displaying these tables, Anca uses jQuery Tables.
Understand your metrics: what are you trying to measure? You can’t measure everything.
Business Intelligence Takeaways
- Learn where the data is.
- Learn a little SQL.
- Make your report actionable.
- Leverage third-party services.
Business Intelligence Tools
Domo business cloud lets you aggregate your data and report on it. (It’s a data warehouse, a term Anca hasn’t used for a long time..)
Google Sheets (or Excel) for viewing all that table output and creating graphs. Sonja says that the number one business intelligence tool is Excel.
If you start using Salesforce, you will need to learn how to use SQL queries.
Points to Remember
- The thing about data is that your brain will impose patterns where they don’t exist.
- In order to turn data into information, you have to understand it.
- Test your assumptions. Figure out ways to validate what you think you’re seeing.
- You can’t just install a tool–if you don’t understand what you’re doing, they won’t be helpful.
- If you do nothing else, download and install AMR Users (assuming your site has more than 2 users)
Where to Learn More
There are online courses and books about becoming a data scientist. For example: