Fax machines. Floppy disks. Landlines.
The list of objects that technology has made obsolete within the last decade is long. Digital technology company AdCellerant wants to add one more item to the list: digital marketing proposals.
The mission of the Denver-based adtech company is to bring quality digital marketing to every business. Hundreds of channel sales partners rely on AdCellerant to make digital advertising sales and fulfillment easier and more efficient.
Engineers, sales and product professionals at the company are currently working on an application that encapsulates all aspects involved in launching a successful advertising campaign, including drafting a sales proposal, turning that into an insertion order, tracking campaign metrics and analyzing ongoing performance. The Proposal Tool, which lives under the company’s UI.marketing umbrella, integrates sales, order and reporting processes in a single user interface.
And while Proposals 2.0, as it’s internally called, is fitted with bells and whistles, it ultimately serves to unify — and in the process, simplify — the multi-step process to selling, launching and running a digital ad campaign.
“Our channel sales partners used to have to work in multiple systems and multiple platforms, which is daunting. Speed to market is brutal and costs are high,” COO and Co-founder Shelby Carlson said. “This tool blends about four different systems into one.”
By speeding up the process, AdCellerant partners and sellers can spend more time in front of advertisers and prospects, increasing retention and profitability.
Below, Jeff Fairley and James Humphrey, senior directors of engineering and technology, respectively, take Built In Colorado through the tech stack behind the multifaceted application. In doing so, they addressed the need for modernization.
“A lot of what is underneath this new proposal tool is completely microservice-driven,” Humphrey said. “We wanted to build a good foundation by which we could start productizing the data that even our own internal apps sit on top of for external APIs.”
The problem with traditional digital marketing
Tell me a little bit about your new proposal application within UI.marketing. What problem does it solve within your industry?
Shelby Carlson, COO and co-founder: We knew we had to redo our proposal tool so that it could encompass the 60+ products that are available to our partners today. Our job is to make Madison-level advertising available to Main Street. So UI.marketing is part education tool, part efficiency tool, as well as a proposal builder that can easily turn into an insertion order with the click of a button.
Our new tool is really for the salesperson more than anyone else. They can educate themselves about what the product does while they’re building it. The hardest part about selling digital is knowing which products work best for both the individual goals and objectives of the client, as well as the industry they’re in.
The tool is inspired and informed by the thousands and thousands of small business campaigns.’’
Can you give me a few examples of some of the products within the suite that this tool offers?
Carlson: The tool gives sellers access to what we call child and parent products. Parent products could include programmatic, video or search. Then, we layer the child products, including geo-fencing, Over The Top or Local SEO. Next, we can match up available targeting, features and benefits to provide a roadmap for sellers to easily build their proposal.
Many small business owners purchase their marketing through a digital sales representative of a media company. We are the operational back end for those different media companies across the country. The end user will be able to have a super sophisticated proposal based on the goals and objectives of their individual wants and needs. And the tool is inspired and informed by the thousands and thousands of small business campaigns we run every month.
What kind of census data are salespeople able to see using the tool?
Carlson: When it comes to geography, imagine being able to say, “This particular zip code highly indexes for a household income of blank.” You can start narrowing down who you’re actually supposed to be pointing your campaign to so that when you are building your data elements for digital marketing, you can leverage data in a smart way.
For example, their advertiser may say, “I really want to go after redheaded left-handed golfers.” This tool allows sales reps to say, “Let’s talk about this demographic and make smart decisions based on the goals and objectives of your campaign.”
Through the tool, it only takes about five minutes to build a very sophisticated proposal. That really provides the value — not only to our partners, sales reps and media companies — but ultimately to the end user, who in our case is the advertiser that we run the campaign for.
The AdCellerant solution
What does the UI.marketing proposal tool look like from the engineering side?
Jeff Fairley, Senior Director of Engineering: Whenever I think about the new proposals tool, the first thing I think of is the change in feature set that we had to worry about and the improvements in performance that we had to tackle. We decided to rebuild the tool from scratch with new tech, thinking about what the users needed.
James Humphrey, Senior Director of Technology: As Jeff mentioned, it started off as a small project. But because of some technology challenges that we had with our initial proposal tool, we decided to revamp our technology that this proposal tool would sit on top of.
We initially were a monolithic application. And we still have that legacy part of our application, which a lot of our customers use. But we’re transitioning from that monolith to a more microservice-oriented architecture. So a lot of what is underneath this new proposals tool is completely microservice-driven and completely API-driven. And a lot of that decision-making came from wanting to be more API-driven, not just for proposals, and not just for optimization and performance purposes. We wanted to build a good foundation by which we could start productizing the data that even our own internal apps sit on top of for external APIs.
You mentioned some technical technology challenges within the initial proposal tool. What were those challenges?
Humphrey: Our monolith had a lot of problems. Not only were we just doing too much within it, we had these really deep verticals of functionality, proposals and reports. We recognized that it didn’t make sense to have a big monolith dealing with all these deep levels of functionality. If a bunch of reports were getting worked on, that would cause load on proposals, and vice versa. We had a lot of load and a lot of performance problems impacting user experience.
For the new tool, we decided on a more service-oriented architecture. So we started cleaving off big slices of functionality into a more microservice-oriented approach.
How does the tool allow the team to utilize this product to the best of its ability?
Humphrey: Our data pipeline is built on top of a Lambda architecture. Lambda architecture consists of two different forks, batch and streaming, that stem from one-to-many input sources. Our primary input is our batch-processing side where data is coming in and batched nightly. That side of our architecture is built, primarily, on top of Java, and our Java microservices use message buses to move data across our pipelines. Individual microservices within those pipelines publish messages and consume messages and all that architecture runs in Google Cloud Platform (GCP), which is a cloud-based architecture analogous to AWS.
We’re very RESTfully-driven now. That’s a decision we made based on what we were seeing with our legacy monolith app. We want to be able to productize that data if we ever want to give anybody outside of the company access to our data as well.
Fairley: As our application runs, we access the data that James’s team maintains through those REST APIs. The REST APIs serve both the internal applications as well as soon-to-be third-party integrators. If someone doesn’t like our reporting dashboard, and they want to pay for access so that they can have their own reporting dashboard, they can build their own front end and use the same REST APIs that our internal teams leverage.
I’ve found in the past that if you don't stay on top of front-end technologies, you can get left in the dark so quickly that it becomes impossible to catch up. Our applications are built on the latest stable versions of React, GraphQL, and Node with TypeScript, and we have a process for keeping dependencies up to date. Every time new releases come out, we’re notified, and we’re vigilant about keeping our application up to date so that we don’t fall behind.
The future of proposals
Where do you see this product headed in six months or a year?
Carlson: On the product side, we’ve taken seven years of manual operation process and automated it. The pre-sales process, the sales process, the insertion order, and the reporting takes most of our media companies three, four or five different platforms to execute upon.
Bringing that into one single toolkit is kind of the end-all-be-all, and we hope to accomplish that full cycle by the end of Q2 of next year.