In this post, we’ll tackle DIDs from our sample call center work list, found here. Doesn’t sound like a big deal, does it? Read on…
What are DIDs?
DID stands for “Direct Inward Dial”. For our purposes, consider this to be the number(s) that the customer calls to reach your center. If you want more background, and crave technical accuracy, see the Wikipedia article.
Within Amazon Connect, the DID is referred to as a “Dialed Number”.
DID I Just Do That?
Amazon Connect supports multiple inbound numbers. You can grab up to 10 numbers from any of the geographic regions supported by Amazon Connect. If you need more, you can contact Amazon. They can supply toll free numbers. They also support the ability to transfer a number that you have today to Amazon Connect.
Best practices state that you want a single number for your customers to access your assisted services. So, why do call centers have multiple inbound numbers? Here’s a SUBSET of reasons:
- To provide premium treatment for high value customers (a “platinum line”)
- To provide direct access to different services from the same company (Sales line, Support line)
- To provide service in multiple languages without a “Press 1 for English, press 2 for Spanish” option
- To provide access for customers that cannot use the toll free line (toll free works in US only, access for customers outside the US)
- To appear local to your customers (offer a local number for each major area that your customers are in)
- To measure marketing initiatives (provide different inbound numbers for each marketing program, measure the inbound traffic)
- To provide an inbound transfer point from another call center system
- To provide an IVR bypass number for your internal operations
- To provide service for another company under their brand (not only outsourcers do this)
- An old number is still out there, and you would lose business if you get rid of it.
Clearly, each of these reasons has implications on how you build up your IVR. Most centers supporting more than one function will have a combination of these reasons. Hey, I’ve seen operations with 100s of DIDs.
Amazon Connect supports the ability to provide different treatment on different DIDs. The DID checking is good, and there’s nothing that I can think of that would restrict my checks. You can branch based on DID, you have a range of conditions you can check (exact match, contains, starts/ends with, etc.) and you can check many DIDs in a single block. I tried 20 DIDs in one block and there is no issue.
Here’s our sample DID detection logic. Pretty straightforward1. Match the DID and head off to the next flow.
What about that Lambda block? It catches stray numbers that are in use, but don’t match any of our DIDs in our DID logic. It logs them, and then sends the customer to a general queue (always be kind to the customer).
After the 20th person has done the 1000th modification, I guarantee that you’ll have at least a few strays. And, if you’re replacing an existing IVR with Amazon Connect (or, indeed, replacing any call center system with another), one of the major challenges is just determining what all your existing numbers are used for, or even if some of them are still in use. This gives you a mechanism to hunt them down, and get them cleaned up early on.
DID I Warn You?
This will be my last DID play on words, promise.
Matching the DIDs is straightforward. What follows isn’t, necessarily.
100 DIDs could begat 100 contact flows. In practice, it won’t be that many but there will be more flows than you first thought. Now, this is not a consideration only for Amazon Connect – it is an issue for all call centers and systems of any size.
The Fish Are Biting
I said at the outset of this series that I only would let you know if there are fish in the lake. I guess there will be a few fishing lessons as well.
1Error handling omitted for clarity
Standard Disclaimer – My comments are based on Amazon Connect at the time of posting, and I endeavor to be accurate. The nature of SaaS is that there can be improvements and changes at any time. If I’ve got something wrong, please let me know, and I’ll correct.