Category Archives: Uncategorized

10 Things Never To Do at a Business Dinner

Business travelers spent $260B in 2012 with food services being the #1 source of expense.  Salespeople love dinners with customers and prospects. Marketers love networking dinners.  We have customer advisory board dinners, pre-board dinners, awards dinners, relationship-building dinners, team dinners, customer appreciation dinners, partner summit dinners, project completion dinners, analyst dinners, investor dinners, … the list goes on and on.

Because business dinners can be so powerful, I am a huge fan of them as a marketing tool.  However, I’ve also been a part of many “dining accidents” over the years — the most infamous being the “white burgundy and stone crab incident” at Estiatorio Milos — almost invariably due to a combination of lack of focus on the business goals of the meal, lack of pragmatism, and lack of adaptation to changing circumstances.

As a result of these experiences, I have composed this list of 10 things never to do at a business dinner.

10.  Lack clear goals.  Whether we’re organizing a 1-1 meal for the CEO and a key customer or a 56-person customer appreciation dinner, everyone on the team should understand the goals for the meal.  Every member of the team should understand why they are there and what they are supposed to do.

9.  Eat in a noisy restaurant.  A universal purpose of a business dinner is for people to get know each other.  That is not going to happen when it’s loud, especially if your guests are a bit older.  Some restaurants are just incredibly noisy (e.g., Wolfgang’s on Park Avenue with a parabolic tile ceiling).  Sometimes private rooms can be quite loud as well, especially if they are not really cut off from the main room.  Avoid live music at all costs.  Avoid low ceilings.  Beware converted bank vaults and train stations.  I’ve seen more business dinners die on this hill than any other.  Fun or hip doesn’t matter if people can’t hear each other.

8.  Have tables bigger than 8.  If people are going to get acquainted, they need both a quiet environment and a table small enough so everyone can hear everyone else.  One friend has a rule that if you want one conversation at a table, then you should limit table size to six.   I think you can go up to 8, provided your team members know there is supposed to be one conversation.  Avoid rectangular tables which greatly limit the number of people with whom one can speak.

7.  Bring too many people.  One advantage of clear goals is that they help in deciding the guest list.  If the goal is to recognize the hard work of a 24-person team, great:  go get 3 tables of 8.  If the goal is for the CEO to build a relationship with another CEO, then either hold a 1-1 dinner or a 2-2, where each CEO brings a lieutenant.  But don’t say you’re having a CEO relationship dinner and bring your sales VP, sales director, account manager, and CFL rep.  It ends up like dating with an audience.  Don’t invite people just because they are in town — you can easily unbalance a dinner by bringing 9 of us and 3 of them, turning it into a multi-conversation, intra-company event to which a few customers are invited.  When in doubt, say no.

6.  Mis-level the crowd.  I think the most important part of networking dinners is that each participant feels like he/she gets value from meeting every other participant.  So if you’re hosting a 16-person CMO dinner, make sure the invitations are non-transferable so you can say “no” when several CMOs want to send their advertising or PR directors at the last minute.  While your PR director may enjoy having dinner with a bunch of CMOs, it’s unlikely to work in reverse.  The worst case is when two CMOs show up and are surrounded by 14 PR directors:  your intended target ends up feeling out-of-place.  It is far better to have 6 CMOs when you were hoping for 16 than it is to have 6 CMOs and 10 PR directors.  Burn that into your brain.  Tattoo it to your wrist.  Don’t not prioritize filling up seats at the cost of mis-levelling the dinner and destroying the event concept.  You can build on a great 6-person CMO event in the future.  You are dead when you host a mis-leveled event.

5.  Leave seating to chance.  Since we’re investing peoples’ valuable time (and probably $100 to $200 per head) in the event, we shouldn’t leave anything to chance.  For larger events, use place cards.  For smaller events, pre-brief the team on who to direct where.  It’s a disaster, for example, when at a square table, you place the two people you want talking next to each other, instead of across.  Make sure it doesn’t happen.  (And if it does, change it per rule 2 below.)

4.  Take more than 2 hours. Business dinners are business.  If you want to add a social part, go the bar afterwards for drinks.  It’s very awkward to leave a business dinner in progress and someone could  end up missing their train and getting in trouble with a spouse, because they expected a business dinner and you ran a lingering social event that took 3.5 hours.  In general, the more senior the invitee, the more likely the dinner is “just another calendar slot” as opposed to a social opportunity.  So when having dinner for 4-6 people at a restaurant (and I’m not in Europe), I tell the waiter in advance that my goal is to be done in two hours and that we want to have two courses and possibly dessert — no shared calamari pre-appetizer, no extra-salad (i.e., salad plus appetizer) shoved in as they love to do at Morton’s.  Just an appetizer per person, a main course, and when the time comes, a decision about dessert.  If things start to go too slowly, have some pre-appointed to leave the table, speak to the waiter discretely and say “get it moving.”  On dessert, if asked first, my answer is, “no thanks, just an espresso.”  If the customer  subsequently orders then I can always join in afterwards. Overall, by respecting your guest’s time, you increase the odds they will say yes the next time you invite them out.

3.  Order very expensive wine.  Here are a few things that can go wrong when you do:  [1] the wine is bad and you end up distracted with the whole rejection and re-tasting process, [2] the attendee is subject to a company policy where he/she has to pay his part of the meal (e.g., government, journalists) so you backfire screw them on their expense report, [3] people love it and you drink three bottles, tripling an expensive proposition, [4] you look pretentious, [5] your company looks wasteful and poorly controlled, [5] your three employees drink it but the customer subsequently announces he doesn’t drink wine and you end up treating the crew and not your customer.  When I lived in France, our classiest sales VP had a simple rule: order Sancerre.  It’s neither too cheap, nor too expensive.  It comes in dry, aromatic white (sauvignon blanc), mild-bodied red, and even rosé (both pinot noir based) so most people will like it.  I’ll demo the Sancerre principle on the wine list from the tony Village Pub, one of the best restaurants in Silicon Valley, where a Corton-Charlemagne will set you back $400 and a Kistler single-vineyard chardonnay $250.  The Sancerre weighs in about $130.

2.  Not roll with the punches.  Entertaining is always full of surprises and you need to roll with them.  We once arrived at The Triomphe in NYC only to find ourselves literally surrounded by a loud, drunken, office Holiday Party.  On arriving, we knew we were dead, so we dispatched a team member to find a quieter spot and did about 3 blocks away.  If a snowstorm wipes out 30% of your attendees, you better eliminate some tables and redo your place settings.  The key thing to remember in rolling with the punches is how to preserve the original goals of the meal.  Twice, I’ve been in cases where 4-5 employees had gathered at a very expensive restaurant (e.g., Morimoto) waiting for a group from a customer who never showed up.  In this case, rolling with the punches should mean eating somewhere else because the company shouldn’t be dropping Morimoto-style dollars on a basic mid-week traveling dinner.

1.  Order the tasting menu.  There are four problems with tasting menus:  they are expensive, they take the whole table hostage because they are ordered on an all-or-nothing basis, they take a long time to serve, and they don’t fill you up. The thing I hate most about tasting menu is not the first check — the $900 check for 4 — it’s the second check, the one for $100 for sliders and wings at the sports bar afterwards.  I am so opposed to tasting menus on business dinners that I actually try to avoid restaurants that offer them; I try to reduce the odds to zero that one person, typically a new employee, will provoke the chain reaction that results in the whole table ordering one.  I’ll do a tasting menu at a business dinner only if we are a small group of known foodies who will order the wine pairings, take three and a half hours on the meal, greatly enjoy it, and not run to McDonald’s right after.  Otherwise, stay away.

I could add as “rule 0” don’t get drunk, but frankly I’ve not seen that rule broken terribly often at the business dinners I’ve attended.  More often, I see it broken at company events — which is a whole different blog post.

I hope you find these rules, and the thinking behind them, helpful to you in optimizing all your business dinners.

Bon Appétit!

I’m Now An Enterprise Irregular

Just a quick post to announce that I’ve joined the diverse group of practitioners, consultants, investors, journalists, analysts, tech executives, and full-time bloggers — known as the Enterprise Irregulars — who share a common passion for enterprise technology and its application to business in the 21st century.

I’ve been a big fan / reader of the Enterprise Irregulars blog and tweetstream for years and think there’s nothing like it in enterprise software.

I was quite honored to be asked to join the group and am very happy to be on board.  In so doing, I’m reconnecting with many old friends and colleagues:

  • Anshu Sharma, with whom I worked at Salesforce.
  • Evangelos Simoudis, with whom I currently work at Host Analytics where he sits on the board of directors.
  • Dennis Moore, with whom I worked at Ingres … we’re talking all the way back to like 1987 or so.
  • Esteban Kolsky, a customer strategist who followed us when I was at Salesforce on the Service Cloud.
  • John Taschek, with whom I worked at Salesforce and who I knew in his journalistic life prior.
  • Jeff Nolan, who I’ve met in and about the Valley and with whom I’ve had some lovely Brunello di Montalcino.
  • Jevon McDonald, who I met during the GoInstant acquisition at Salesforce.
  • Merv Adrian, a fellow data aficionado and someone who I’ve known since his days at Giga.
  • Paul Greenberg, a CRM author and expert … and ditto on the Brunello
  • Ramana Rao, with whom we partnered when he was at Inxight and I was running MarkLogic.
  • Sameer Patel, who I’ve met in and about the Valley.
  • Zoli Erdos, publisher of Enterprise Irregulars and Cloud Ave, and who I’ve met in and about the Valley.

And those are just the folks I already know in some way!  The full list of Enterprise Irregulars is here.

Now all I have to do is to finish the labor-of-love that I’ve been writing on SaaS renewals rates.  I should be done in about a week.


Video on Why Customers Are Moving EPM Systems to the Cloud

Here’s a three-minute video of me talking about why customers are choosing to move their EPM systems from traditional on-premises systems like Hyperion and onto cloud-based systems like Host Analytics.

Sales is from Mars and Engineering is from Venus

I was talking to a startup CEO the other day and he said:

Lately my VP of Sales and my VP of  Engineering have been at each other’s throats.  Badly.  I need to write a book called Sales is From Mars, Engineering is From Venus or such.   Any ideas for me on how to manage this conflict?

Sure.  Two things typically go wrong in this relationship.  One is that you have a strategic alignment problem.  The other is that you’re treating sales and engineering the same way and — to your point — they are different animals and should be treated differently.

First, you must get this conflict back in control because it sounds dysfunctional, is hurting your organization, and to paraphrase Machiavelli, warring Princes means a weak King.  Since you’re the King, you need to end this war, posthaste.

The Strategic Problem with Sales and Engineering:  Alignment

Sales/Engineering tension typically comes from a lack of alignment around strategy.  Sales, by default, wants to sell anything to anybody and it’s up to the CEO to make sure that sales is ring-fenced enough into a target market that they can’t keep generating effectively random product requirements and be taken seriously.

Towards that end, you should create — as an executive team — a one-page document entitled “Our Target Market” that describes, using terse bullets, what the perfect prospect looks like when he or she walks through the door.

The more a given customer looks like that ideal, the more their voice should be heard in the product requirements process.  And conversely.  This helps create the notion of strategic vs. opportunistic revenue.  The former is revenue coming from the target, the latter is revenue that you will take, but you will not modify your product or strategy for it.

I have seen terrible battles between Sales and Engineering and this lack of alignment, and the long-list of fairly random product “deficiencies” that accompany it, is usually the cause.

Avoid the blame game.  Your e-staff is one team and you win or lose together.  If they’re fighting, it’s either because they’re bad folks and need replacement or they’re good folks and they are not aligned on the mission.

The Communication Problem with Sales and Engineering

Founder / technical CEOs love to reason.  They are reluctant to bark orders.  Instead they prefer to lay out options and debate merits, eventually arising at consensus around a strategy.  Logically and dispassionately.

That style tends to work well with Engineers (and marketers).  It tends to work far less well with Sales.  At the same time you might be thinking, “Gosh, I’m doing such a great job reasoning with my sales VP,” he or she is probably thinking “when is this guy going to make a stinking decision and tell me the answer?”

Sales are soldiers.  They like to be told to take hills and they will fight hard to do so.  They trust that you have chosen the right hill to take. They see you as the General leading the army.  They want you to do your job and they will do theirs.

My advice with Sales is to stop reasoning with them.  Explain quickly why you are sending them on the mission.  But don’t reason forever.  Make calls decisively and give the order that they are awaiting.  They will see you as a stronger leader and respect you for such.


  • Reason with Engineering
  • Command Sales

It may sound harsh, but in the end, I am certain that if you fix these two issues your Sales/Engineering conflict with disappear.

Forbes: Will Host Analytics Do Unto Hyperion as Salesforce Did Unto Siebel?

I’m happy to report that Forbes Magazine has written a great Forbes Insight Profile on Host Analytics and me.  The story is entitled Will Dave Kellogg’s Host Analytics Do To Hyperion What Salesforce Did To Siebel?

Here’s an excerpt:

“We’re trying to do to Hyperion what Salesforce did to Siebel and the entire CRM category.  We’re a cloud disruption company,” says Host Analytics CEO Dave Kellogg.  Host Analytics is using a cloud-based delivery model to disrupt the traditional EPM (Enterprise Performance Management) market by offering the improved execution speed, broader organizational accountability, and deeper business insight of modern enterprise EPM applications.

Host Analytics cloud-based delivery model extends these benefits with the business agility, improved security, and cost advantages of the Cloud.  EPM is a fancy acronym for the applications required by the Finance department to manage the accounting and budget planning process.

The whole story is here.

How Much Does My General Manager Need to Know About My Department?

Question from a reader:

Dear Dave,

How much does my general manager need to know about my department?  Frankly, he/she doesn’t seem that interested in what we’re doing or spend that much time with us — which I suppose is a good sign in some ways.  On one hand, I’d love to deep dive into all the great things we are doing.  On the other, I don’t want to bore him/her with a lot of unnecessary detail and be seen simply as yet-another engineering manager who doesn’t “get” the business.

The first answer to this question reminds of an 800-pound gorilla joke.  Q:  Where does an 800-pound gorilla sit?  A:  Wherever it wants to.

That is, your GM needs to know whatever he/she wants to know about your department.  So I’d listen closely about what he/she requests in terms of information and make sure you deliver that.  Don’t make the all-too-common mistake of not providing what is asked for, but offering lots of other information instead.  That’s like serving more icing than cake and, personally, I see it as either incompetence or a deliberate diversion.

Because (I happen to know that) your GM is not a functional expert in your department, he/she may not have a clear sense of what information is of interest.  So he/she may place the burden on you of figuring out what information is appropriate, say, for a 30-60 minute quarterly ops review presentation, or a 1-1 meeting.

In that situation, we can use some first principles to guestimate what the GM wants to see.  All good GMs care about:

  • Sales, which is typically measured by ACV in a SaaS company.  Remember, this is the raison d’etre for the GM role — i.e., the reason the company has a unit GM is precisely because they want someone focused on and accountable for the unit’s sales.  Because sales is a trailing indicator, GMs are usually focused upstream on pipeline as a leading indicator of sales.
  • Customer satisfaction as measured directly through surveys, anecdotally through escalations, and indirectly through renewals.
  • The team.  Who are the top performers and how are we retaining them?  Who are the bottom performers and how we are improving their performance?  Who is in the middle and how we are developing them?
  • Objectives and accountability for achieving them.  What objectives are you working towards, why were they chosen over alternatives, how do they tie to the unit and company goals, and how are you doing on achieving them?  Personally, I am a stickler presenting prior-period objectives in a verbatim form, because too many people try to hide the very normal tendency to miss some of them.  Far better, in my opinion, to reach a bit in setting objectives and miss a few than game the system so that you are always hitting 100%.
  • Key departmental metrics.  I think every manager should have 5-7 topline metrics that they use consistently to measure the department over time.  Because they will be used over time as the company scales, it is often best to normalize these.  Cases/agent or cases/customer is more interesting over time than just the number of cases.  Bugs/customer or bugs/engineer is more interesting than total bugs outstanding.

So let’s try to apply the above to answering your question.

  1. I’d sit down with the GM and determine what your topline metrics should be — this is often a long and interesting series of conversations.  This exercise is, in my experience, never as simple as it might appear.
  2. I’d sit down with the GM and ensure you are aligned on quarterly objectives.  In a matrixed organization where you are reporting to both a product line GM and a functional vice president, this can be harder than meets the eye.
  3. I’d think hard about how what your department does ties to sales.   While your functional boss will focus on about  best practices, your GM will — as noted — invariably focus on sales.  For development that might be about neutralizing competitive features.  For QA/QE, that might be producing a higher-quality product on faster cycles than the competition.  For UE, that might be studies reflect not only on how designs work with existing customers, but particularly on how prospective customers react on an first-impression basis.  For techdocs, that might be how can the documentation make the product sufficiently better so as to influence sales, perhaps by highlighting competitive strengths in examples and tutorials.
  4. I’d think hard about how your department ties to customer satisfaction and renewals, using the same approach as above.
  5. I’d provide a quarterly update on the team citing top/bottom performers, key hires and informally discussing team issues within the quarter.
  6. Finally, I’d make a point to make myself “more than a department head” by spending as much time as possible learning the product / market / competition and talking live with customers, e.g., during corporate visits, user conferences / groups, et cetera.  This will facilitate all the prior points and help  you get ready for your next promotion!

The Independently Wealthy Salesperson

Technical founders and entrepreneurs can easily overlook the coin-operated nature of salespeople. Why? Because they aren’t salespeople, they’re product people and they’re just not wired the same way.

Founders might be motivated by changing the market, popularizing a product, or just proving they are right. Salespeople, almost all the time, are motivated by their compensation plans, so the compensation plan should be the de facto expression of what you want them to do and how they should spend their time.  Ergo, as mentioned in this post, you should get them done before kickoff. And put a lot of thought into them.

Why? Because a typical salesperson will spend the whole weekend after you give them their comp plan in Excel modeling how much money they will make under various scenarios. I’d also say to make them as simple as possible both to make it clear what you want salespeople to do and to avoid the inevitable unintended consequences that often accompany complexity.

One huge question is whether comp plans should be capped. Almost all salespeople would say no. Part of the reason they’re playing the game – particularly at a startup – is for the lottery ticket.  Think: while I know my on-target earnings are $250K, I want to have a shot at earning $1 or $2M – that’s what drives me to work those killer hours.  So while I recommend leaving comp plans uncapped, I also recommend that management model the full range of scenarios and, for example, accelerate rates in the 100-250% of plan range but to greatly decelerate them after that.

You can always hedge your bets in the compensation plan terms and conditions (e.g., plan can be changed at any time to correct for errors or unforeseen circumstances), but if you actually use that language the whole salesforce will know and you will quickly lose the lottery-ticket value in your comp plan. It is far better to put some more thought into the plan on the front-end. I know one guy at a startup who did a $10M deal off a $1M quota and received only a fraction of his stated comp. Because he didn’t want to burn bridges, he just quit. But in this scenario, everybody loses.

Outliers, however, can take several forms. I know one sales manager who groups salespeople into three buckets:

  • Those who clearly understand their comp plans. They sell what’s incented, when it’s incented, and make the most money per sales dollar.
  • Those who mostly understand their comp plans. Those who do a good job following the plan incentives, but not a perfect one.
  • The “independently wealthy” who seem to pay no regard to the incentives in the comp plan

I love bucketing reps in this way both because it’s funny and it immediately prompts an important question. Why are these reps not following the plan? Perhaps it’s just sloppiness or stupidity. Or perhaps there is more going on. My advice is to analyze reps in this way, show them that if they had sold different products at different times how their pay would have varied and then ask them why they didn’t. While some people invariably just miss the point, you might also discover “good reasons” why your people aren’t following your plan: maybe it’s too complicated and they don’t understand it, maybe they don’t think the higher incentive offsets the additional risk of selling a new, strategic product.

Or maybe they truly are independently wealthy and just doing sales for fun. But I doubt it.