The ability to see, understand and respond to consumer actions in near-real time is a fundamental necessity for today’s marketers. To engage in relevant conversation with the consumer, you need to act at the speed of the consumer. If you can’t do that, then your messages may become unappealing and irrelevant. When that happens, even your most loyal consumer relationships can be usurped by relevant real-time communications from your competitors.

A study from Harvard Business Review with global C-suite executives found that more than two-thirds of respondents increased spending on real-time customer analytics solutions over the past year. And not a moment too soon. Consumers want their interactions with brands to be relevant, meaningful and consistent across channels. To achieve this, the consumer experience needs to be elevated to the front and center of planning, and the market realizes this. A study from Gartner suggests 89 percent of marketers will primarily compete on customer experience instead of price in the coming years.

Read More: How to Create a Winning Marketing Operations Playbook in 3 Steps

To deliver exceptional customer experience, every marketer must have a real-time understanding of the consumer’s motivation, intent and expectations. With this information, they can create an emotional connection between their brand and the consumer. There are other components involved such as creative and content, but without real-time information, marketers are stuck in a world of batch communication and inferred interest.

Modern marketers see the transformational possibilities being able to calculate new offers and decisions in real time brings to customer experience. Marketers in retail, banks, telcos and media have shown that more personalized and contextualized real-time messages perform better than batch campaigns.

For example, consider how dependent cable and internet providers are on their call centers to manage customer interactions. Call centers are expensive, and this has led to considerable investment and development of chat interfaces. The ability of chat conversations to occur and resolve customer concerns based on real-time understanding is transformative. Think how beneficial it would be for the consumer experience if the chat agent knew there was a service outage where the consumer is based, that the consumer called earlier but the call was dropped due to the network issues, and that the consumer also looked on the provider’s website for package upgrades within the past 24 hours.

In this example, real-time decision making would help the chat agent to proactively apologize for the dropped call, inform the consumer of the outage issue, give them an estimate for resolution and then offer a discount on a package upgrade for the inconvenience. The consumer gets a connected omnichannel experience with a positive outcome, and the cable provider gets an upgraded relationship for less expense. This type of immediate analysis can boost a company’s bottom line.

Read More: Matching Marketing to the Modern Consumer: 5 Ways to Rethink Qualitative Research

So, how do you get started with real-time customer analytics? Adding real-time capabilities to existing customer analytics may require new technology and staff, but perhaps most importantly you’ll need to revamp processes. Here are a few suggestions for approaches:

• Begin with one project for which the scope is well-defined and build the application vertically before deploying horizontally. The lessons learned on that initial project will help you avoid making the same mistakes on a broader, more visible playing field.

An example would be to connect data from two channels together, analyze it and feed the insights back into one of those channels for execution. In my experience, combining web and email to then customize web content is an easy starting point. At the danger of oversimplifying it, this can be done using a single tag across both mediums to create a consistent and structured data set. Expose that data to your analytics engine, then have the analytical recommendations connect to your content management system. This should drive higher engagement on your website.

• Start small and simple. Make your mistakes and learn your lessons on a small stage, not on Broadway.

If we continue with the previous example, you could start with one small section of your website and one small email campaign focused on a specific product or opportunity. Don’t try to overhaul your entire content and email in one go. Make a mistake on 5,000 emails, not 1 million. If your analytics are done correctly, you should be able to scale the results.

Read More: 7 Sure Shot Content Marketing Trends

• Remember that it’s not a disaster if the initial project doesn’t succeed on the first attempt. All it means is that you have another shot to make it work. When it’s done correctly the next time, the project is a success, and the next order of business is to scale up. Once the initial project succeeds, it is also much easier to sell the next real-time analytics project.

The adage of “right message, right person, right time” is now “right message, right person, right now.” Modern marketers must be able to do this. Marketers that forge the way with real-time customer analytics will be ahead of the curve competing on customer experience.