Hello, Enterprise. Goodbye, CPM.

alexkelleher“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Alex Kelleher, founder and CEO at Cognitive Match.

Advertising technology is mostly paid for on a usage basis: serve more ads, make more money. This has been true of media buying, ad optimization, ad verification and so on — all adding cents or dollars to an overall CPM cost for showing an ad.

But is this sustainable? Or will the downward pressure on per-thousand pricing, and the recognition that quality, not quantity, is what advertisers want to pay for, force a shift to a longer-term enterprise model?

Enterprise or Software-as-a-Service (SaaS) pricing means that I charge you a fixed fee per month or year to use my product. Ideally, that locks you into a long-term contract, which allows me over time to maximize my profits. You get the benefit of using my software as much as you like for the same price. In addition, I’ll charge you for value-added services instead of trying to wrap some service into the CPM I charge you or enforcing minimums, upfront charges and so forth.

This model is the future of ad tech for five key reasons.

1. Platforms, not black boxes

Companies such as Rocket Fuel, Turn, The Trade Desk and Quantcast are not trying to deliver a black box to meet their advertisers’ end goals of showing some ads and making some sales. They are building platforms that allow agencies and advertisers to transparently manage spending across all channels, including mobile, social, video, Web and search. The game is competing with Google’s self-serve platform and getting product embedded into their customers. That’s where the long-term value lies.

2. Product = product + service

In building companies that maximize value for shareholders, scalability is always seen as key. The idea of building out a company with more people is less attractive than building out a customer base with a product you endlessly replicate. However, most buyers of advertising technology will take a long time to learn how to self-serve, and therefore will need help. Help equals people’s time, which the ad tech companies then have to pay for. So they wrap that cost into the overall CPM, which is difficult to sell.

Enterprise software addressed this a long time ago by pricing product and service separately, and making product a predictable, fixed fee over a long period of time.  Companies from IBM to Omniture make billions this way, and still scale very nicely.

3. CPMs aren’t going up

CPMs, particularly outside of the sexy new mobile and social areas, aren’t exactly soaring. Commodity does that: As things get cheaper to deliver, people expect to pay less for them. Only charging for ads that are viewable, verified and not served to robots or fraudsters has the same downward pressure — fewer ads are shown with greater effect. This creates an immediate conflict of interest. As an ad tech provider charging CPMs, I would want to do everything I can to increase ads shown to achieve the same advertiser outcome. SaaS pricing allows that pressure to fall away, so providers and advertisers can focus on getting the job done for the least reasonable price. Budgets can then be spent on things like service, making up the lost revenue for providers and boosting their profitability. That’s a win.

4. Venture capital loves contracted revenues

Most of Terence Kawaja’s logo salad of startup companies in the ad tech space is fueled by venture capital. Venture capitalists invest for one reason: to return funds to their LPs. This makes them perversely risk-averse. The concept of an insertion order with a customer that can be canceled within 48 hours leaves them with sleepless nights. What they really want to do is go to bed, safe in the knowledge that their startup has a contract that will take months, or years to cancel. They can easily value such businesses based on forward cash flows, and then they can sell them for much healthier multiples. So they are now desperately seeking places to put their investments that have this, because CPM for a new piece of the ad stack just doesn’t do it for them anymore.

5. The stock market loves contracted revenues

We’re back in a period of frothy IPOs and a rush to go public. Guess what the markets love, just like VCs. Certainty of revenues and a clean and understandable business model. Rather than media buys, which excite by their size but then disappoint with margins, SaaS models generally are very clear as to the profit built in and how they scale.

Will the change happen overnight? Of course not, and some of the ad business will stay the same because media buying makes a lot of sense on a CPM basis. But everything else — the layers of technology that end up with showing an ad in that media buy — is positioned for pricing as a licensed platform and service fees, which are essentially consulting fees. This will liberate ad tech companies from the stress of IO-to-IO business, and help venture capitalists sleep at night.

Follow Alex Kelleher (@alexkelleher), Cognitive Match (@CognitiveMatch) and AdExchanger (@adexchanger) on Twitter.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Enterprise or SaaS pricing is actually probably the ideal for ad tech companies for the many points you listed but I would venture that the reason it hasn’t taken hold is how the advertising ecosystem is setup.

    Agencies generally don’t have technology budgets to invest in stacks. They are a service-based business that could get dropped by their own clients at any time and that relies primarily on passing on costs to the clients (sometimes at a markup). If they can’t distinguish or line-item a technology expense to each client, then they can’t pass on the costs, driving their own costs up. If they can get dropped at any time, they aren’t so willing to invest in expensive enterprise stacks over a year.

    Now the flexibility of SaaS with some month-to-month pricing could solve a lot of this, but given the set up time and training often involved, it becomes harder and harder for ad tech companies to recoup their investment if customers could themselves leave. It just becomes harder though growing with the customer without setting tiers based on usage and then defining which dimensions to measure usage. CPM is just the most understood. Even if you were to make it “impressions managed a month,” that would be pretty volatile for most agencies. You could set up so it’s campaigns, but there is little value in archiving past campaigns for clients the way that archiving contacts for a CRM makes sense and therefore drives CRM pricing. You could set it up as users, but there isn’t that many people in a given agency that would use the tool actually. So clearly there are challenges in SaaS pricing for Ad Tech and hence most people go with a minimum + CPM (minimums allow the firms to count it as predictable revenue).

    CPMs are also used by ad tech vendors because they do have variable costs that scale with delivery of service on an impression basis. Unlike most SaaS or enterprise solutions which don’t necessarily have incremental impression costs, ad tech would then need to bake in expected impression costs into each tier (which would mean making assumptions or restrictions on campaigns sizes and # of campaigns per tier).

    Now, where these points break down are if you are talking about selling ad tech direct to brands/advertisers either as an ad tech vendor or as an agency. Brands/advertisers do have technology budgets and can commit but this assumes they are moving agency functions in house (as some are). Some consulting firms (like IBM and others) are in fact doing what you’re saying and selling the ad tech as an enterprise solution with services bundled around it which is perhaps what agencies need to start doing before the consulting firms (who are launching ‘agencies’ or marketing consulting practices) eat their lunch.

  2. Thanks Victor – great points! You’re right any change will take time, but it feels like the tipping point is close now, at least for some parts of the industry. One to watch….

  3. On #2, eventually the product IS the service. The more capable and flexible the platform, the more I as an advertiser, can drive my effective CPM toward zero while improving overall results. The downside is I probably can’t figure out how to use the platform(s) all perfectly. I don’t have the staff nor the time to train them.

    Give me the Salesforce model every time–dirt-cheap, highly flexible tech platform with lots of hooks to other platforms.

    I’ll bet on Ad Tech companies that have the best professional services capabilities in the long run.

  4. I agree that CPM (and % commissions) are not good models for ad tech.

    However, if productivity is a big part of your value proposition, a SaaS subscription model is at odds with the value your ad tech delivers. The more productivity you deliver, the fewer people needed to do the job and the fewer seat licenses you sell (or minutes used). So, you make less money for delivering more value. Conversely, you make more money for delivering less value.

    Would transaction fees that vary with value delivered be a better model?

  5. Mark – great comment, I agree that services are undervalued (in all senses) in ad tech companies. There’s still a very strong role for us humans…

    Joe – thanks for the comment, and the tweets. Fits with Mark’s comment in the sense that the added value services/product extensions is where the profit can be made, and the scale of taking on lots of customers at a relatively affordable monthly fee.