Anis Jaffer: Hello everyone, welcome to Tuesday Talks, a live discussion series where we shed light and bring truth to emerging topics in the communications industry. I’m your host, Anis Jaffer, Chief Product Officer of Numeracle and I'll be co-hosting today’s session with Frank Pettinato, CEO of Avantive Solutions. It's good to have you, Frank, welcome.
Frank Pettinato: Thanks, Anis, it’s great to be here.
Anis Jaffer: For today’s episode, we’ll be focusing on Rich Call Data, which we will be referring to as RCD. RCD enables enterprises to present brand identity such as brand logo, name, and call reason on consumer devices.
Building solutions on RCD alongside Numeracle has been Frank and his team at Avantive. For those of you who don't know much about Avantive yet, they are a full-service digital marketing and contact center solutions provider utilizing new and emerging technologies such as RCD. Their innovative approach helps their clients maximize campaign success, improve customer engagement, and ROI.
As soon as RCD became available we worked with Frank and his team to build the architecture and launch branded calling for an early adopter client and here we are to share the story. Frank, to start, can you tell us why it was so important to be the first out of the gate to make branded calling available to your customer brands?
Frank Pettinato: That’s a great question, Anis. At Avantive we try hard to drive value for our clients through innovation. I define innovation as solving real business problems through the creative use of technology. Our clients look to us to drive innovation and rich call data allowed us to test the value.
Honestly, in researching it there was very limited real data available other than marketing material and what I call “brochure wear.” There was very little hard data so we partnered with Numeracle and one of our clients to develop a fairly robust and extensive POC. It ran over several months to prove the ROI, understand the challenges of deploying the technology at scale, assess the real impact to the outbound calling environment, this program, in particular, that we tested was an outbound, understanding customer reaction, and more importantly, pulling data out through some type of managerial reports.
Anis Jaffer: At this point, it's important to highlight some subtle differences in the solutions that are available right now in the market. There is Verified Calls by Google, which renders rich call data such as logos on Google-enabled devices like Pixels. There are also branded call solutions available on mobile carriers through their analytic partners. They're usually referred to as branded calls but there is some variance within that. One is where you have the full experience where you have the logo and call reason that is presented.
Then you have just the brand name as part of the solution; it's like CNAM but a better version of CNAM because you get to control the name and it also provides uniformity across the platform that provides it. The important thing here is this RCD is considered out-of-band to STIR/SHAKEN, meaning, regardless of whether the call is STIR/SHAKEN enabled or not, you can send the original data over the data network, and hence it’s out-of-band. For sending RCD as part of STIR/SHAKEN some standards have been presented. Adding the RCD PASSporT as part of the delegated cert, or entity cert, is the most widely discussed. However, we don't have those solutions rolled out yet.
For today's discussion, we will focus on what's available today. So, Frank, what was the experience deploying the solutions that are available like Google Verified and branded calls by carriers?
Frank Pettinato: Happy to speak to that. For clarity, the examples that I’ll be talking about today are on an outbound or OTM use case but we are in the process of looking at some inbound/return call use cases. Because of the high volume of OTM, it felt like a better strategy to apply. What we ended up doing is testing the data and looking at options we had. We worked with your team to develop the technical solutions and negotiated with the different third parties that provide RCD, like Google and two of the three other major mobile carriers.
We began looking at how to put this together in a unified solution but more importantly being able to capture the data for further analysis. As you mentioned, one of the things that we also were interested in testing was the RCD PASSporT standard, which has now been set by ATIS, but right now that has been adopted by many carriers. Once we gained agreement from the brand we tested we confirmed the brand name to be presented, which, as you mentioned, may be different than the CNAM, it gives you some flexibility. We selected a logo and the discussed call reason and began to deploy the solution among these different platforms.
This deployment had many different aspects to it including using API, providing origination caller IDs to the third party carriers are, providing relevant brand data and brand approvals, as you know, there's some sensitivity of using specific brands so all these third parties would want to make sure we, as a third party, would have brand approvals. The investment here was a significant amount of time and resources, but overall, it went relatively smooth considering this is new technology.
Anis Jaffer: We also had some lessons learned, especially implementing the solution, like Google Verified Caller where we had to implement APIs. All this new technology is great but we have to deliver results; that’s where the rubber meets the road. Solution providers claim that branded call display, whether it was just a brand name or logo, added a lot of value for call originators and enterprises. I've seen numbers as high as 50% increase in contact rates, but in general, around 30% is what I've seen quoted. What has your real-world experience been?
Frank Pettinato: We didn’t know what to expect, we read the general material available but we modeled a neutral live connect rate in the absence of any hard data suspecting that if people knew who was calling and were truly interested in listening to the brand would pick up and others would not.
Honestly, that was our starting point, but after about a couple million calls we saw a pattern emerge and we found that it was driven by the propensity of the record being called. So no surprise that high propensity records saw a big improvement, a low propensity record saw a neutral and, in some cases, small decline. This isn't necessarily bad, in my opinion, because the people who did answer ended up converting at a higher rate, upwards of 25%, higher in some cases. Ultimately, this reduced the cost per sale and had a very positive ROI on the RCD cost. Low propensity programs struggled in some cases, coming in at neutral or slightly negative ROI. It's about making sure that you have the right offer at the right time.
The way to think about this is the brand experience, in our case, is in a roulette game in the sense that we're not just hoping that whoever answers the phone will be a buyer. By presenting a brand and a logo and perhaps a call reason, it smooths the interaction out and targets the brand loyalists who are willing to hear the offer. But in some cases, the offer is still not right for them but it leaves us an opportunity to do that.
Generally, we’d say to plan for about a neutral from an OTM standpoint, but the people who you do reach are probably going to be very interested in hearing your message.
Anis Jaffer: So what I can infer is, branded calling, especially when it's done right, improves confidence in which the consumer or the client you're trying to reach will answer the call. It also gives them the confidence to give a callback. Call completion is not the only metric to track, you also have to look at other things including call back, call duration, and quality of the call. Are there any other metrics that you were looking at?
Frank Pettinato: We saw an increase in callbacks due to the brand, in some cases with the logo being left in the call history. So someone may have missed a call but rather than it being a random phone number, they'd understood that it was perhaps a brand that they had an association with. Additionally, since we were connecting with people who are interested in hearing from the brand, in many cases, as I mentioned, the conversion rate was higher. That caused an increase in an average call length because we were engaging with the customer and converting versus agents waiting and dialing.
The quality of the calls was generally stronger and the anecdotal feedback from the agents was that the call started on a much more positive note as the callee, or the person we were calling, knew the brand that was calling and therefore the initial objections were overcome. The call was set up to be more successful, additionally, we pulled some of the initial NPS data showing a market improvement. That the Net Promoter Score of folks that did engage with a brand through branded calling increased. We still need to evaluate that but early indications on that promoter score are very very positive, further supporting our position that the brand loyalists were the people interested in hearing from the brand and answering, and perhaps people who are not, are choosing not to answer.
The unexpected results were fascinating. After a couple million calls we had a significant amount of data. We looked at the call recordings through our voice analytics platform looking to see if we can get any indication of if the person being called had mentioned the fact that they saw a brand or a logo. I was delighted to hear that we didn’t see that. The customer or prospect answered simply subconsciously. They realized the call was legitimate and so they answered so it wasn’t a gimmick, which could evaporate over time. I, for one, was very pleased that the customers didn’t mention the branded call but the data showed the results I just mentioned.
Anis Jaffer: That’s very interesting, that’s a fascinating observation. Presenting the branded logo, in effect, is not the reason making people answer or call back. It has more to do with trust. Sure presenting a logo is new and it’s cool because it’s a novelty thing, which could fade, but subconsciously, if you trust that a call with a logo and a verified sign, then you’re more likely to call back or answer the call. That’s what you’re inferring, correct?
Frank Pettinato: Absolutely, I think it’s all about that trust factor that’s gained based on who is calling.
Anis Jaffer: So we’ve talked about improvements and we’ve talked about how you’ve been able to leverage this, so let’s shift into some of the challenges for those in the audience who are considering testing this before they fully implement. What are some of the challenges you faced? Can you share some thoughts?
Frank Pettinato: With the work with Numeracle, the technical challenge has been overcome so the challenges became collecting and interpreting the results. We added the carrier append so we know exactly which carrier was calling on each call to our business intelligence platform to help analyze the results.
We were able to sort performance by mobile carrier for which ones were receiving rich call data and which were not and referencing it against historical data for performance. We had to collect and analyze NPS data related to rich call data versus not. We had to review millions of calls through the voice analytics platform to get a sense of agent performance and callee sentiment and emotion.
We also had to understand program by program nuances and differences. Just because we have rich call data and get a live answer doesn’t mean we have the right offer at the right time, which impacts ultimate sales results. So we wanted to normalize for that, and that’s what I meant when I mentioned high propensity and low propensity.
Optimizing call reasons to help drive the best outcomes, perhaps being overly descriptive or under descriptive on call reasons may drive callee behavior. We dedicated a lot of time and resources out of our operational excellence team to develop a business model, which we believe is now set up and can evolve as additional data is collected.
The bottom line is there’s definitely good ROI for some programs, but perhaps not all. The ability to analyze the results and justify the ROI is critical. You may find that you’re not presenting the right offer at the right time to the right audience and the RCD is not necessarily going to solve that.
Anis Jaffer: So as a contact center, you still need to watch the call results and analyze the data, so RCD is not going to solve everything, but you can leverage it to further enhance your chance of success. With that background, how do you see the cost versus benefit in implementing these solutions? As we’ve found as we’ve gotten this enabled, pricing is varied depending on the solution provider. What are your thoughts on that?
Frank Pettinato: We continue to look at the ROI and looking at different factors and results but I, for one, was very surprised at the range of costs that were presented from the various third-party providers. This indicates, no surprise, that there’s a lack of maturity in the market and some level of market inefficiency due to how new it is.
We started by developing our own models and asking where we need to be to make this viable based on some assumptions. We then worked with the providers and with Numeracle to try to get to a price point that was tolerable for us, especially during these early stages because we had limited information. Overall, our initial thinking came in very close to the ROI we expected, which was great news.
Over time I expect the market to become a bit more efficient and for pricing to become more ubiquitous as the technology and the reporting becomes more standardized. I also mentioned earlier that I’m very interested in the RCD PASSporT in STIR/SHAKEN where the standard is singular and can be universally adopted by all the carriers, which we think may improve the cost structure.
Anis Jaffer: I agree with you 100%. We are all working towards the end goal of making rich call data as part of every call that we make, especially if that can be added as the RCD PASSporT to the secure layer of the STIR/SHAKEN certificate. That would be ideal. We are still quite a ways away from there but we are getting there, so that’s the goal.
And when consumers receive the call they should be able to trust a logo that's presented. There has to be some kind of verification upfront so that the trust layer is still there. Otherwise, we’ll end up where we are with CNAM today where we can’t trust the name that shows up, in part because there are different databases and all kinds of data is in there. But by providing the verification layer upfront and then validating that, attaching it to the certificate, or sending our-of-band helps in enabling that trust.
Before we open this up for questions, Frank, do you have any parting thoughts if you were to summarize a few items for the audience? What would those be?
Frank Pettinato: I think the branded calls establish trust by confirming the identity of the businesses in real-time. The users or the people being called have the confidence of knowing who is calling and trust that the calls are not being spoofed. I believe the branded experience drives live answers from brand loyalists setting us up for a better and more successful call experience for the agents and the callee.
The combination of trust and brand experience has driven higher conversion rates, sales performance, and potentially stronger NPS. Those are the three takeaways I feel confident that these past 3 or 4 months we've been working at this have shown.
Anis Jaffer: Great thanks, Frank, that was an excellent summary. At this point let’s answer some questions.
Molly Weis: We do have some questions. The first one is: Why has it taken so long to get RCD rolled out? Is this related to the recent STIR/SHAKEN deadline?
Anis Jaffer: So RCD, as we said before, has a couple of different ways to be implemented. Currently, as it is available, let’s take Verified Calls by Google as an example, it’s out-of-band. You create the data, which is the logo and the call reason which you push to the Google servers, and then when the call is presented, they pick up and it is shown. Similarly, some of the analytics partners have similar solutions that they’ve rolled out and these are also out-of-band.
With STIR/SHAKEN, that is supposed to solve the authentication for the number, not the data. So the STIR/SHAKEN rollout is targeted at securing the telephone call and adding RCD is another component to it. It's an additional element that would go on top of STIR/SHAKEN, but as of today, the standards have been published but the adoption is not there yet. But then you have out-of-band, which are over the data network solutions that are already available.
The RCD standard itself came out of IETF. Now, the STIR/SHAKEN and the RCD PASSporT that has been discussed to be added as part of STIR/SHAKEN is within ATIS, which is the Telecommunications Industry Group. But IETF, initially, ruled out RCD as a spec and that's been available for quite some time now, it’s just that the Telecom Carrier Association is now getting ready to implement that standard on STIR/SHAKEN.
Molly Weis: Let's go to our second question, which is a multi-part question: Is RCD expensive to implement? If so, how will this affect call delivery for brands who are not able to display branded content to enhance their presence like others who can? Will this create a new standard of who we trust?
Frank Pettinato: ‘Expensive’ goes back to two parts. One is by looking at it as a true financial ROI. So if you are financially performing better after RCD vs. before. Or you are looking at it as another way to deliver a lot of brand impressions, which by definition has some value. So depending on the brand, they may choose to use RCD to deliver more impressions of my brand and the side benefit is they can potentially have some improved performance. That would be a brand-by-brand experience. To call it expensive, you’d have to take a hard look, but we justified a very positive ROI so we are comfortable with saying that it is a financially viable investment.
Delivery of brands who are not able to display content versus those that can I do believe that over time the consumer will recognize that there are just random numbers calling and they don’t know who it is. The level of interest in answering them will reduce as you start getting those verification flags or ‘Caller Verified’ from STIR/SHAKEN, and now rich call data, where at a minimum you’re getting the brand name, up to including a logo or other graphical responses including a call reason. I do believe that over time, though I don’t know how quickly, the folks that do invest on this will start to make the impression with the consumer on a phone call will resonate with their loyal customers and their customers. Brands that choose not to adopt this may end up seeing a reduction in live answer rates because of that trust factor.
Anis Jaffer: Just to add a bit on that, I’m not sure I would term it as expensive, it comes down to the use case. But there is a variance in the pricing that’s available, which could be expensive depending on which solution it is.
What was the third part of that question?
Molly Weis: The third part was: Will this create a new standard of who we trust? Which I think Frank addressed.
We have one more: What are the biggest hurdles to adoption? What expectations should organizations interested in RCD have about the timeline to availability?
Anis Jaffer: Today, the way to implement, if you were to go directly and work with each of those providers, that's going to be different. You would have to call a different API, you would have to set it up in three or four different ways. That, to me, is quite a hurdle, which is why Numeracle is trying to solve that problem so that enterprises do not have to implement it in four different ways. That’s one hurdle.
The other hurdle is that the pricing itself is varied so there's no uniform pricing yet. I'm hoping that eventually, we'll get there. If you are trying to justify ROI it's a little bit challenging because of the variance in pricing.
Frank Pettinato: I think the hurdle to adoption is doing an ROI analysis, the technology piece has been worked out. For the expectations of the timeline, we've gotten two of the three major carriers in production, as well as Google, and the third major carrier we hope to have up and running by Q4. By the time the year closes out, we should have fairly ubiquitous coverage by all the carriers.
Molly Weis: If I understand this correctly, an enterprise could go directly to each branding provider, so what was your consideration in working through Numeracle as being a single provider?
Frank Pettinato: We work with Numeracle on many different things, specifically our reputation management. They had a relationship already in place with Google, so that worked out nicely with the API. They had been negotiating and working with the other three third parties providing the service, so those relationships were very valuable in getting in contact with the right people to implement it. Ultimately, there are some nuances on what the solution the Numeracle’s putting together that is a value-add. We felt that the solution at Numeracle was going to deliver to us and our client was a holistic one and not necessarily using the best talents of our team to cut and cobble something together that would be less than that adequate. Ultimately, it was ease and expertise as the way I would describe it.
Anis Jaffer: Thanks, Frank, that was some great live feedback on Numeracle, I appreciate the feedback on that.
I would like to thank you for joining, Frank, it was awesome. I'm hoping that we continue to work on this and maybe have you back for another session in a few months when we get more data and more to talk about. For the rest of the folks, our next session is going to be on August 10th when Rebekah is back. We hope to see you there, thank you for joining today!
