Our client since 2017, Here Comes The Guide is a women-owned & operated website showcasing top-notch wedding venues and down-to-earth planning advice with a hint of sass.
We’re proud that they’ve been our client for many years. We’re not proud that it took so long to speed up their slow site search. Previously most any search took multiple seconds to complete.
HCTG is a directory of wedding resources, and its most important job is to help couples and wedding planners quickly and easily find what they are looking for. Slow search means high friction. Friction is terrible on a website.
After evaluating possible technical solutions and building a simple demo to test our assumptions, we settled on integrating Meilisearch into herecomestheguide.com.
We brought the idea of solving the search problem with an appliance like Algolia or Meilisearch to the table. We educated the client, conducted the discovery, built the prototype, and ultimately completed the full integration and launch.
Faceted site search on HCTG has been reduced from multiple seconds to mere fractions of a second. It’s so fast it almost looks fake!
Here Comes The Guide is a woman-owned and operated business that provides a directory of wedding venues and vendors for engaged couples and wedding planners in the U.S. The site runs on Craft CMS and previously used Craft’s native search capability. Significant performance gains had been realized through effective caching strategies. However, due to the vast number of search combinations across the many different attributes of wedding venues, no caching strategy could capture a large enough sample of search sets. A new solution for search was needed, and would have to be built from the ground up.
The Solspace process is to continually improve website reliability by reducing website friction. The customers of a website should be able to come into, through, and out of a site with the least resistance possible, finding what they need and completing a transaction as quickly and easily as possible. We helped our clients at HCTG think in these terms. The process added even more emphasis on what we and our clients already knew: fixing the speed of search was a first-order priority.
The slow search was friction. It was a barrier to website users achieving their goals. This flawed user experience made our clients vulnerable to their competition.
Sites like HCTG provide their users with what’s called faceted search. This means that users can sort through large and complex data sets across multiple attributes and criteria, narrowing and refining their results as they go. For example, a couple coming to HCTG might be in search of wedding venues in Oregon, in the Bend or Portland areas, specifically those with mountain views, barns, or lofts with in-house catering. As soon as they see the results, they may change some of those filters to increase or decrease the search results.
This faceted exploration should be fluid and frictionless. They should see results change as rapidly as they explore their tastes, interests, and available options.
Imagine our couple is using herecomestheguide.com to explore their ideas for a wedding at a winery. They are slicing and dicing results as they move through their options. They trigger one search and as they wait for it to load, and wait and wait, they get distracted and into a conversation about their dog, who is barking at the door. The next thing you know, the page has finally loaded with their results and they have gone off to walk their dog. We lost user engagement. They didn't see their winery. Their winery didn't get contacted by them. And the whole point of them visiting the website was lost.
But now that search is split-second fast. Our couple can find their dream winery instantly, before the dog barks. That winery gets contacted by them promptly. They book the date and collect their deposit. The couple is happy with the results, and the whole business model of the website is validated in seconds.
We started by looking at Algolia as a possible solution to our search problems. Algolia pioneered this new wave of high-speed, high-relevance search on custom data sets. We spoke with the Algolia sales staff, discussed the specifics regarding our client’s problem, but ultimately weren’t able to get to a place where we could reliably predict our clients' costs. This was a deal breaker. A responsible business needs to be able to forecast expenses. We couldn't with Algolia.
Meilisearch is also a ‘search as a service’ solution like Algolia in that it has a hosted offering. This means you can sign up for an account and they will take care of everything, scaling and all. Additionally Meilisearch has open-sourced their offering. Anyone can set up a Meilisearch server and launch it into production. Meilisearch makes the value apparent, and keeps the pricing clear and transparent. Our client could predict their costs.
Meilisearch also anticipated our needs by providing well-thought-out and well-documented developer utilities. There were significant resources available for our Vue JS approach. The backend API was intuitive and approachable. We could launch local versions of a Meilisearch server on our laptops. The documentation was comprehensive and easy to approach. In short, our needs were anticipated and met every step of the way.
The fact that Meilisearch was so well positioned to support our developer efforts translated into us coming in under budget significantly, delivering much earlier than planned, and beating our time-to-market estimate well ahead of our client's high traffic season. What was initially a source of clear and present business risk was mitigated by excellence in the underlying product and its support.
We experienced no friction and as a result, we were able to provide an extremely high level of Web Reliability to our client.
Our client’s site is well established. It has been around since 1998. As such, the business logic is well-tested and tuned. Any major new change to the site would have to dovetail well into the existing structure, and that's no small task with a site of this size and complexity. That's why a partial headless approach was so helpful.
With a headless approach to web development, you are not relying on the underlying CMS templating, content system, and engineering to compose pages and user experiences. Instead, you decouple front-end code from the backend and rely on a simple API to extract data from the database. The front-end code no longer cares what the ultimate source of data is as long as it's available through a set of known API calls.
Meilisearch uses the power of this kind of decoupled architecture to make integrations into complex websites much more easy and more approachable. We have one component, a plugin we wrote for Craft CMS, that passes data from the CMS up into the Meilisearch server via its standard API. This builds the search indexes. We then have front-end code written in Vue JS that sends search queries to Meilisearch, receives the results, and composes the changes to the page for the user.
Now that we have improved web reliability by reducing friction in the core functionality of the site, we can move to better monetize it. Our client, as with many sites like theirs, monetizes the site by offering featured listings above the fold on key pages. Previously the available business logic for this was constrained. But our Meilisearch implementation has greatly expanded the possibilities. We can now offer much more effectively targeted feature listings to a broader cross-section of HCTG clients.
We are predicting that revenue will increase as site popularity increases due to improved performance. But we also believe that revenue opportunities will continue to expand due to a more open set of paid content targeting capabilities. We are excited to see how Meilisearch will continue to deliver for our client going forward.