12 Nov 2019
Build vs. buy a site search solution – Part II
 Karanvir  Gupta
Karanvir Gupta

Various factors must be considered regarding the market scenario, existing vendors, and identifying our needs and specific gaps. These are before making the decision of Build vs Buy.

In this blog post, we will look deeper at what it means to build a Site Search solution from scratch. This should act as a reality check for all ecommerce businesses before they take a final call.

Site Search solution: An outline

Building a Site Search solution end-to-end is no cakewalk. Multiple components need to be working in tandem with each other for site search to function seamlessly.

First and foremost, you would need a search platform to be developed and ready. It could take 2 years (approx.) to build, test, and make the platform stable and ready to use. Even if you use SOLR as the foundational pillar, you will need resources to upgrade as and when the SOLR platform evolves. SOLR doesn’t come with all the site search features built internally. Those features must be built over the top of the platform and custom-tweaked as per your business requirements. This would require full-time developers.

The main challenges, however, are as follows:

  • Long gestation period to make Site Search go live
  • Investment of time, capital, and manpower in non-core competency (product)
  • Inability to evolve the solution with changing technology and business needs
  • Opportunity Cost involved for the time your search solution is not live

Data and the practive role of AI & ML

Once the search tech stack is up and running, you will need data servers to host and manage the vast amount of data. This would be an ongoing expense, and as your system evolves, the need and expense for servers will keep rising.

However, data alone is not enough. You need to convert this data into meaningful information and useful business insights that can make your search solution efficient. It would need AI and ML models and algorithms in place that can automate the most complex part of the site search solution. This is humanizing your search solution. The ability to show contextually relevant search results to shoppers (or products shoppers are most likely to buy) is built over time.

In essence, it is intelligence that is gathered and put to use on the go. It would need resources in terms of developers and analysts to build and manage this system. The seamless running of this system is key to ensuring that your Site Search works just fine.

The need for Site Search features

It is great that you have AI and ML models ready to power your site search solution. However, to utilize these features, your eCommerce site needs to have site search features available and integrated. These features are listed below. You will be required to build all these features so that your AI and ML models can play the role they are built for.

Some other allied tasks include:

  1. Upload of the product catalog to the system
  2. Feed indexing
  3. Setting up the analytics so that data points can be gathered across the shopping journey
  4. Tracking user flow
  5. Testing analytics integration
  6. Validation and testing of the overall system

Note: These tasks (1-6) will also help you make a choice between various competing vendors offering you the Site Search solution. (in case you are opting for the buy option).

Other key modules include:

  • Merchandising capabilities
  • Analytics and reporting
  • Dashboard/Console for a 360° view of the business

For your site search system to function flawlessly, you would also want it to be capable of performing all of the above along with basic search functionalities. From boosting or burying products, creating a landing page for promotion or holiday campaigns to sorting – merchandising gives you a plethora of opportunities and ways to move your inventory.

You want to be absolutely certain about the search queries that worked, the fast-selling products, trending searches, and the searches made recently while shoppers came to your eCommerce store and searched for and bought products there.

Why is Analytics and Reporting vital?

Analytics and reporting let your ecommerce managers stay on top of all these details. These basic functionalities play a key role in bringing out the best performance from your Site Search solution. Because unless your site search solution helps you make informed business decisions, it is as good as nothing.

Now, you can decide if you want to build these functionalities on your own and build the site search. Or, you can outsource to a vendor offering you all of it. I just listed all the tasks that need to be performed for your Site Search system to be up and running.

Each and every task has an associated:

  • The cost involved: Some of the action items above are one-time expenditures, while others are recurring expenses that will keep happening month-on-month
  • The time involved: You will have to check how much time it takes to build the site search system from scratch. This will compare to the time the vendor takes to go live with the product.

Importance of timeline

Last but not least comes the timeline for this project. Are you ready to wait for 2 years for the Site Search solution to be built in-house? Can your business wait and bear the opportunity cost associated with it? How quickly do you want the solution to go live?

Final note

The answers to all these questions will help you make a decision between building this product in-house and outsourcing this to a vendor specialized in offering Site Search services to many ecommerce stores like you.

The decision is yours!

In case you are looking to buy a Site Search solution for your ecommerce store, book a demo!