OraNA :: Oracle News Aggregator

Got a lot of positive feedback on my post from yesterday about some lessons learned while growing OctetString, so will write a few more articles along these lines...

Given that consulting billing rates and hours go down in tough times, many consultants will undoubtedly decide to build businesses to "product-ize" some of their solutions.

This is completely possible -- our own Oracle Virtual Directory started out that way, as mentioned yesterday. Some of our best partners were started this way as well.

However, there are a number of common traps that consultants fall into when they enter the software business.

1. Repeatability, not complexity...

Repeatability in software gives you the ability to scale your customer base.

Consultants often work in the role of car mechanic -- look under the hood, find scary problems, suggest some solutions that will require parts and labor, and finally plan and implement that solution. Each customer has a new problem requiring different parts, plans, and execution.

Software is a very different business. You're looking for as much commonality as possible between customers so that what is delivered can be repeated at other customers with the minimum possible effort.

This doesn't mean that you can't solve complex problems or require services to implement. It simply means that the product-ized part of your product shouldn't be different for every customer.

It also goes without saying that building your software on standards-based middleware will help reduce the amount of post-sales time spent doing customized integration with each customer.

2. Don't design around one big customer...

Even companies that don't originate from consultants tend to fall into this trap a lot, but consultants do it almost every time because they tend to be solving a problem that they encountered at a particular customer.

It is wonderful finally finding a customer or prospect who will spend a significant time helping you understand their requirements. You should certainly listen to them -- they are the customer, eh?

The trick is to use that customer to validate your approach rather than try to solve every esoteric problem that the customer might have through your software.

You might consider providing extension points, allow for customer designed templates, and so forth to accommodate their needs without building less repeatable stuff into the product. You'll also want to consider that your product may not be the right place to solve a particular issue.

3. Consulting isn't Software Sales

Many consultants have great people skills. They can help customers understand complex technology as it relates to the customer's own environment. Customers often base decisions about technology purchases in-part on recommendations from expert consultants.

However, this doesn't always (or even often) translate into being able to actually sell software. Not that you can't learn to do so, but the process of selling is very different from the process of pitching a solution as a consultant.

As a consultant, you have high credibility in part due to your independence. As a vendor, that credibility is diluted to some degree, even when you're still trying to help the customer do the right thing to solve their problems.

The overall process goes far beyond what you say and how credible you are with a customer. You're going to need to educate yourself and bring in the right people to help you be successful.

What happens when you don't know what you're doing?

A (now) funny story from our very first sales call (with Fannie Mae, oddly enough) back in early 2001 before we hired our first sales person or took the time to better understand the sales process:

I was on the phone with the customer while another person from the company was acting as the "account manager". Nobody is in the same room. The customer's first question: "What does this solution cost?" Oops! We hadn't priced it yet and hadn't discussed how we would handle this question. After an uncomfortable silence, the customer's question was answered (after a flurry of background instant messaging).

Thankfully we got better at this with time, brought in people that had experience selling enterprise software, and things worked out well.

I should also point out that the very first customer we did end up selling to was BEA Systems. This is why a portion of Oracle Virtual Directory's original 1.0 release is actually embedded in every copy of WebLogic 7.0 and above.

Many of you know that I came to Oracle through the acquisition of OctetString. You may not realize that I co-founded OctetString in early 2001, which was during the last downturn. In fact, we were negotiating our first software sale when the 9/11 terrorist attacks occurred.

So I read with great interest Jason Calacanis's email (and blog post) discussing how startups can better survive an economic downturn. Given that he too started his last company (WebLogs, acquired by AOL--think Engadget) during the last downturn, he's got very solid advice. I don't agree with a few specific items (never had a reason to test dedication with Sunday morning meetings), but overall a great read.

I thought I'd share a bit of advice and a few tales from that same period of time, but in the enterprise software space.

When we first started OctetString and created what is now the Oracle Virtual Directory, we had a number of pre-baked customers that were lined up to buy our software. Unfortunately, most of these were telco customers and by mid-2001 our phone calls weren't simply finding people who had been pink-slipped, but entire divisions that had been abandoned and certainly weren't going to be buying software from us anytime soon.

What kept things going was pretty simple:

1. Keep costs low -- especially recurring

While all expenses should be reviewed, you're going to want to pay particular attention to things that are recurring, including people, rent, etc...

When we needed hardware, I just made a trip to the local liquidator. HP-UX server: $800. Solaris box: $995. A bit like going to a junk yard and not as glamorous as handing over a check to your local rep, but it works.

Until almost 2003 we didn't even have an office, and even then I just used a Regus facility in order to share some common services with other companies (and not worry about anything related to maintaining the office itself). Not to mention the lease was relatively short-term (and I loved their coffee machine).

2. Retain an insanely dedicated core group

Jason makes this point and mentions seeing who shows up for a Sunday morning meeting or such to see who's dedicated. I think you'll know who the right people are even without having to test them.

These people won't care about fancy offices or silly perks. My first sales guy closed a critical deal with Pfizer in the United economy class line at O'Hare Airport on the way to Germany. Another critical deal with Coca-cola was closed in a phone booth at University of Illinois while a kid was breaking up with his girlfriend in the next booth (it's not you, it's me). Not having an office or business class for international trips didn't seem to stand in the way of his performance.

Others were equally (and probably more) dedicated. In the earlier dry times it wasn't uncommon to be deferring paychecks.

3. Make something people actually NEED -- particularly during bad times

When we started the company, we noticed that most enterprises were still doing multi-year projects to consolidate and synchronize all of their user repositories with a technology called "meta-directory".

The underlying need was to get all user information into a single place for portals, ERP, HCM, CRM, and related business applications. These are big, important applications and every one of them needs information about usernames, passwords, roles, department numbers, reporting hierarchy, and so-forth to function.

This may seem trivial, and today using software like Oracle Identity Management it's a lot easier, but at the time it was a black art requiring lots of consultants, lots of software, and Big Dig style project timelines (and success rates, unfortunately).

We simply shrank many of these projects from years to days and the results couldn't be ignored.

One particular customer implementing a CRM solution with a lot of consultants estimated that they saved something like $10m in consulting over-runs alone.

You have to be having a huge impact that can't be ignored simply because you're not the right vendor. This is especially true when times are tight and customers become more conservative. Customers know that a lot of smaller vendors won't make it and don't want to be stuck with abandonware.

Looking forward to comments. Thinking to do a few more posts on this topic if there's any demand.