The build or buy question runs through the heart of technology as well as through the heart of every business leader.
Even the largest companies don’t do everything in-house. Apple doesn’t make its iPhones, Foxconn does. Businesses outsource everything from public relations to human resources to website development and design.
Programmatic advertising has added some nuance to this age-old build-or-buy conundrum. These days, it’s possible to build your own programmatic trading desk. Some, including Sprint, Dr. Pepper, Snapple, Amazon and Netflix, have done so successfully. Whether it makes sense for your business depends on your tech resources and staffing, your appetite for big projects, your willingness to make costly mistakes, and your desire for control.
Before you take a leap in either direction, let’s consider the arguments for and against each approach.
Building programmatic: High costs, but more control
The main benefit of building your own programmatic platform is control. An in-house, custom-built programmatic platform ensures complete control over your buying process. When buying a platform, you’re just one of several advertisers for which the vendor will optimize and potentially manage programmatic spend.
Lack of control means that the ads you’re paying for might not go where you want. Think car ads designed for California customers that accidentally run in the Massachusetts market or an extra zero on the price of an item in your catalog causing it to show up in ads over and over, driving up CPC costs to absurd levels. Marketers with custom-built programmatic solutions can keep a handle on what products their ads promote, what sites those ads show up on, and can maintain control of their customer data, a prized asset these days.
Whether building your own programmatic platform makes sense for you depends on your tech resources and staffing, your appetite for big projects, your willingness to make costly mistakes, and your desire for control.
But that control comes at a cost. Even if you “lease” an off-the shelf bidder — which is just one part of many technologies you’ll need on your journey to build your own solution — it will cost at least $10,000-$30,000 a month, depending on your monthly traffic and how many auctions you engage with. Even more costly will be the 10-person at a minimum team of engineers and data scientists. In addition to the RTB (real-time bidding) bidder and 10 full-time salaries, consider that a full-fledged programmatic desk would need the following technologies:
- A data management platform replete with user profiles to power targeting and optimization
- Optimization algorithms for deciding when and how much to bid
- Creative server, template design and item/recommendations
- A user interface for reporting and management of campaigns
- A streaming data pipeline to power the DMP, optimization and reporting
Because OpenRTB is an open standard, it may seem like setting up a programmatic platform is easy. It’s not. It’s a very involved project with the all the usual complications and cost overruns.
That said, if you build your own system and can match the performance of existing solutions you will probably save money in the long run. However, the first part of that equation is often overlooked by marketers that build their own solution. Imagine Apple spending the time, resources and money to build a Formula One race car and while it does get around the track fast it’s only half as fast as a Ferrari.
Buying programmatic: Slightly less control, but far fewer headaches
A few years ago, if you wanted the benefits of a custom in-house built solution you had to build your own. The alternatives were agency-style managed service providers that executed programmatic buys on your behalf. However, you can now buy a software platform that gives you most, if not all, of the benefits of building your own solution with a lot fewer of the headaches of going it alone.
As with any SaaS (software-as-a-service) product, if you become too dependent on it then you run the risk that the vendor will raise prices. You’ll also be at the mercy of the vendor when it comes to which features are prioritized for development.
If you choose the right vendor, there’s a middle ground in which you can use vendor tools that act very much like a custom-built solution.
But these concerns can be somewhat alleviated if the vendor tailors it’s solutions for specific types of advertisers. A platform that includes the laundry list of technologies outlined above can offer a similar level of data control and customization as a custom-built platform. There’s usually a limit to how much a SaaS vendor will let you customize your ads and product feeds, but that’s not the case with every vendor.
In that sense, build-versus-buy is a false dichotomy. If you choose the right vendor, there’s a middle ground in which you can use vendor tools that act very much like a custom-built solution.
Buying programmatic is the better choice for most
If you’re a company with the budget, software engineers and data scientists to build your own programmatic trading desks from scratch, congratulations. But those resources are probably better spent on more pressing company initiatives.
It’s tempting to copy the efforts of giants like Netflix and Amazon. But even those companies could have benefited from buying an in-house solution that aligns with their business goals. They built their programmatic ad platforms because … well … they could, and at the time they had no other viable options. And more power to them. On the flip side, there are household-name ecommerce brands that tried to build their own programmatic platform, came up against massive problems like not enough inventory or scale, and then came running back to buy in-house tools.
For most companies, buying an in-house solution is the optimal choice to get the necessary technology components (bidder, algorithms, DMP, creative server, UI and reporting), the data transparency and the customer support to successfully run their own ad campaigns at scale.
Are you optimizing your advertising toward the only performance metric that truly matters to your bottom line? Here’s everything you need to know about incrementality.
Download the Guide