Is your company still searching for great search?

The intersection of audiences and the content they’re searching for-that is, giving the right content to the right people at the right time-is a delicate balance conducted by multiple teams behind a technology curtain. A user ordering their daily Starbucks feels like their mobile app just gets them — so why isn’t it the same with every digital experience? Just like Starbucks , they use these apps and software every day. So why don’t those interactions count for something?

The thing is, making something effortless is often a lot of hard work behind the scenes. And this decision can be compounded when your existing search platform reaches its end of life, like Adobe Search and Promote has.

A recent survey of technology professionals conducted by Coveo found that 81% stated that search had only increased in importance over the last 12 months, and 85% took this one step further by increasing their budgets for search. The survey also surfaced a number of challenges companies were coming up against when trying to successfully achieve relevance in their digital experiences. 

A chart shows 85% of surveyed companies stated they'd increased investments in enterprise search

In the webinar, Successfully Replacing Adobe Search and Promote, Stephen Rahal, director of product marketing at Coveo, was joined by Eric Immerman, practice director of search and content at Perficient, and Jeffrey Poland, director, web development at Cloudera, to discuss considerations, challenges, and best practices for putting that budgetary resource for search to good use when evaluating replacements for Adobe Search and Promote.

Learn how Coveo augments Adobe Experience CloudBlog: How Adobe Experience Cloud and AI Deliver Best-in-Class Digital Experiences Together

Making Search Relevant, Manually

While search is a deceptively simple concept, it’s certainly not one size fits all. Building search that meets everyone’s needs can require a metric ton of customization. Especially when you start with an open source platform like Solr. 

“This required a lot of development work, a lot of customization,” Poland said, in describing the path Cloudera took that brought them to Adobe Search and Promote. “We were actually building a lot of things into Adobe Experience Manager to provide things like boosting; a lot of things that come out of the box [with Adobe Search and Promote].

“Over time, we realized the amount of time and effort we were putting into it didn’t make sense with the results we were getting out of it.” 

To achieve scalable indexing and search, Cloudera migrated to Adobe Search and Promote. This provided them with an easier interface for business users of all stripes to do what they needed to do.

A screenshot shows how Coveo appears within the Adobe Experience Cloud interface. 

“But you still needed someone to go in and create rules and different things like that to curate the results,” Poland said. “This provided a good situation for us for a while … but we were still stuck in a place where we were [manually] curating it. There wasn’t a good way to take the data that’s coming in from users — telling us how they’re using our tool — and then really implement that in a real way.”

Once they found out Adobe was sunsetting Search and Promote, Poland knew it was time to find a tool that would resolve the manual curation issue. He started looking into Coveo.

“The premise of Coveo was really that not only do you have these tools that Adobe Search and Promote provides,” Poland said, “but you have these machine learning algorithms that can lay over it. It’s tied into analytics and can help search become much more of a data-driven tool.”

But just as with any major purchase, it helps to know tips and tricks up front. Here’s a case study of Cloudera’s journey from Adobe Search and Promote end of life to Coveo, and the four best practices they followed for success.

4 Best Practices for Successful Adobe Search and Promote Replacement

Best Practice 1: Take a Greenfield perspective on relevancy

Starting from scratch can be a terrifying prospect — but sometimes that perspective is needed. One of the best ways to begin a new search project is by looking at and evaluating your current system, but without any of the prior baggage.

“We had a scenario with a customer,” Immerman said, describing a commerce search use case. “They had 14,000 rules in their production system. Everything from typos to synonyms to boosting and campaigns and categories and everything. And they were insistent on keeping all of them. ‘We need to have every single rule, we’ve spent so much time working on this.’”

The sunk cost fallacy was boxing them into keeping their old search experience, even though they were swapping out their search engine.

“We wrote scripts, we migrated everything into Coveo. And they were like, ‘we’re getting the same relevancy that we used to.’”

To get the most out of shifting to a new paradigm like Coveo and machine learning, you have to let go of the crutches and workarounds you’ve had to put in place to make things work. Those same workarounds are now holding your business back. This means letting the machine learning take on the burden of understanding and serving up results to the user.

“Look at your current interface, your current rules, current details like that: we want to make sure that people are taking a step back and, say, ‘let’s operate under the assumption that we have good relevancy out of the box,’” Immerman said. “And by doing that, we end up with a lot better experience than if we try to keep everything that’s moving over.”

“It’s been a huge time saver to not have to think about, okay, how are we going to wrap this into this new platform?” Poland said. “Or how do we write this rule to do this, because we’re already getting the results that we’re expecting. And that’s definitely been very encouraging and move it along much faster than we thought we would.”

Best Practice 2: Establish clear goals for your replacement strategy

Everyone arrives at search replacement having taken a different path, with different requirements, background, and organizational structure to work within. This necessitates defining what goals you want to achieve with your migration. With 50+ connectors, an AI-powered search platform like Coveo can simplify the process.

“Everyone has a different use case, and a different project plan on how you want to do it,” Poland said. “So being clear on where you want to surface content and how you want data to come in can help you evaluate.”

A photo shows a search engineer choosing from a range of content sources

For Cloudera, using the sitemap connector was a simple and quick way to migrate the taxonomy of their digital assets from Adobe Search and Promote to Coveo, due to having metadata embedded in webpages. This allowed them to run Adobe Search and Promote alongside Coveo to compare the two. Eventually, Poland said that Cloudera will run the AEM connector to their dev environment to test out a different implementation.

“I think the biggest thing is just understanding, especially if you’re coming from Adobe Search and Promote, what you’re currently doing,” Poland said. “How are you currently structuring your metadata and your taxonomies? And then think about how you would apply that.”

Best Practice 3: Start with a pilot search project to show time-to-value

When search isn’t a focal point or throughline throughout a company’s various experiences, it easily becomes disjointed. Achieving true relevancy across all touchpoints, regardless of what use case you’re working from, requires the unification of many moving parts—which can be a tall order when replacement time comes.

At Cloudera, there were many teams across the organization that are still using Solr and hadn’t even transitioned to Adobe Search and Promote.

“It can often be a hard thing to get out of [doing a certain level of customization],” Poland said. “A lot of teams were down in the weeds doing their own customizations, and it can be a hard thing to justify the cost on their end. Often, we see teams become siloed and it’s hard to get that governance of saying, ‘hey, we all should move to something.’”

Moving the marketing site from Adobe Search and Promote to Coveo is proof positive that Poland needed to show that bringing machine learning into the search equation can make things easier. He’s hoping to bring additional sources in over time, as knowledge of Coveo pollinates throughout the company.

“What we’ve seen with Coveo is that we’ve almost completely replaced the search for our marketing site,” Poland said, “and gotten anywhere from 95% to 98% of what we had to do custom for Adobe Search and Promote from just out of the box with Coveo.”

While not yet live, Poland said they were interested in getting users interacting with the site and allowing the machine learning to start taking more of the work off their hands. For one example, with Adobe Search and Promote, they’d had to build tons of rules just around synonym detection—something that the Coveo platform can handle autonomously.

And with that autonomy, more business users can interact with the platform, instead of relying on IT to add in new rules to achieve the results users demand.

Want to give Coveo a spin?Trial: Try Coveo search for free

Best Practice 4: Get taxonomy owner buy-in with hands-on education

Change is difficult, even when it’s for the better—it can involve growing pains and discomfort that most would rather stay with the status quo to avoid.

“I’d say once a year at Cloudera, there’s a big initiative to discuss a global unified search,” Poland said. “And the main challenge is just the owner of the taxonomy. What does a term mean—from a marketing standpoint, versus a documentation standpoint, versus knowledge base, etc.

“Getting cohesive agreement that we’re going to tag things in a unified way is important, especially when you do often slip into silos.”

And nothing hurts more than a siloed team that’s short on talent. Coveo’s survey of tech professionals found that 95% were experiencing a talent shortage. One of the ways Poland hoped to encourage Coveo buy-in was the fact that the platform can be autonomous but still user friendly for the non-technical person.

A photo shows a business user checking analytics inside the Coveo platform

“When you have a tool like Coveo, where a lot of tuning can be automated—but on top of that, you can get your [business users] in there and [have them] understand how the platform works,” Poland said. “Then they’re curating and are able to interact with it and see what they can do with it.”

That direct platform interaction and education can spark interest in applying AI-powered search to other applications and systems throughout an organization, creating a groundswell that can bring teams into alignment for unification.

“I don’t know that our Coveo implementation begins and ends with basic search,” Poland said. “The sky’s the limit on what we think we can do with it.”

It’s one thing to fix search—it’s another to evolve it and provide relevance wherever your users are. Machine learning can help companies scale, while relieving the burden on employees and providing a great experience for end users.

Coveo offers a platform that democratizes artificial intelligence for companies and organizations of all sizes, empowering them with AI-powered search.

If you’re interested in hearing the whole conversation, the recording for Successfully Replacing Adobe Search and Promote is available now.

Listen to the webinarWebinar: Successfully Replacing Adobe Search and Promote

Dig Deeper

Wondering where to start when evaluating enterprise search options on the market?

In our Buyers’ Guide for Enterprise Search Platforms, we explore four key components to consider: crawling and indexing content across silos, applying intelligent automation and prediction to search, learning from interactions, detecting intent, and guiding actions, and how to scope your search beyond a single touchpoint.

Get your copy today!

Download your guideEbook: Buyers Guide for Enterprise Search Platforms
Share this story:

About Wayne Applebaum

Dr. Wayne Applebaum is a data scientist, with a background in designing analytics, search and machine learning frameworks that impact a customer’s bottom line. With over 30 years of experience working with C-level executives, he understands how to harness the power of data science to solve business problems.

Read more from this author

About Ashley Garst

Ashley Garst is the Senior Content Editor at Coveo. She has more than a decade of weaving words together and is a ninja editor, slicing extraneous prose like fat from a steak. A California native, when she’s not at work, she’s running in the hills of San Francisco, practicing her creative writing, or playing with her cats.

Read more from this author