Category Archives: demandgen

I’ve Got a Crazy Idea:  How About We Focus on Next-Quarter’s Pipeline?

I’m frankly shocked by how many startups treat pipeline as a monolith.

Sample CMO:  “we’re in great shape because we have a total pipeline of $32M covering a forward-four-quarter (F4Q) sales target of $10M, so 3.2x coverage.  Next slide, please.”

Regardless of your view on the appropriate magic pipeline coverage number (e.g., 2x, 3x, 4x), I’ve got a slew of serious problems with this.  What do I think when someone says this?

“Wait, hang on.  How is that pipeline distributed by quarter?  By stage?  By forecast category?  By salesrep?  You can’t just look at it as a giant lump and declare that you’re in great shape because you have 3x the F4Q coverage.  That’s lazy thinking.  And, by the way, you probably don’t even need 3x  the F4Q target, but you sure as hell need 3x this quarter’s coverage [1] and better be building to start next quarter with 3x as well.  You do understand that sales can starve to death and we can go out of business – the whole time with 3x pipeline coverage — if it’s all pipeline that’s 3 and 4 quarters, out?”

I’ve got a crazy idea.  How about as a first step, we stop looking at annual pipeline [2] and start looking at this-quarter pipeline and, most importantly, next-quarter pipeline?

What people tell me when I say this:  “No, no, Dave.  We can’t do that.  That’s myopic.  You need to look further out.  You can’t drive looking at the hood ornament.  Plus, with a 90-day average sales cycle (ASC) there’s nothing we can do anyway about the short term.  You need to think big picture.”

I then imagine the CMO talking to the head of demandgen:  “Yep, it’s week 1 and we only have 2.1x pipeline coverage.  But with a 90-day sales cycle, there’s nothing we can do.  Looks like we’re going to hit the iceberg.  At least we made our 3x coverage OKR on a rolling basis.  Hey, let’s go grab a flat white.”

I loathe this attitude for several reasons:

  • It’s parochial. The purpose of marketing OKRs is to enable sales to hit sales OKRs.  Who cares if marketing hit its pipeline OKR but sales is nevertheless flying off a cliff?  Marketing just had a poorly chosen OKR.
  • It’s defeatist. If “when the going gets tough, the tough get a flat white” is your motto, you shouldn’t work in startup marketing.
  • It’s wrong. The A in ASC stands for average.  Your average sales cycle.  It’s not your minimum sales cycle.  If your average sales cycle is 90 days [3] then you have lots of deals that close faster than 90 days, so instead of getting a flat white marketing should be focused on finding a bunch of those, pronto [4].

Here’s my crazy idea.  Never look at rolling F4Q pipeline again.  It doesn’t matter.  What you really need to do is start every quarter with 3.0x [5] pipeline.  After all, if you started every quarter with 3.0x pipeline coverage wouldn’t that mean you are teed up for success every quarter?  Instead of focusing on the long-term and hoping the short-term works out, let’s continually focus on the short-term and know the long-term will work out.

This brings to mind Kellogg’s fourth law of startups:  you have to survive short-term in order to exist long-term.

This-Quarter Pipeline
This process starts by looking at the this-quarter (aka, current-quarter) pipeline.  While it’s true that in many companies marketing will have a limited ability to impact the current-quarter pipeline — especially once you’re 5-6 weeks in — you should nevertheless always be looking at current-quarter pipeline and current-quarter pipeline coverage calculated on a to-go basis.  You don’t need 3x the plan number every single week; you need 3x coverage of the to-go number to get to plan.  To-go pipeline coverage provides an indicator of confidence in your forecast (think “just how lucky to do we have to get”) and over time the ratio can be used as an alternative forecasting mechanism [6].

this qtr togo

In the above example, we can see a few interesting patterns.

  • We start the quarter with high coverage, but it quickly becomes clear that’s because the pipeline has not yet been cleaned up. Because salespeople are usually “animals that think in 90-day increments” [7], next quarter is effectively eternity from the point of view of most salesreps, so they tend to dump troubled deals in next-quarter [8] regardless of whether they actually have a next-quarter natural close date.
  • Between weeks 1 and 3, we see $2,250K of current-quarter pipeline vaporize as part of sales’ cleanup. Note that $250K was closed – the best way for dollars to exit the pipeline!  I always do my snapshot pipeline analytics in week 3 to provide enough time for sales to clean up before trying to analyze the data.  (And if it’s not clean by week 3, then you have a different conversation with sales [9].)
  • Going forward, we burn off more pipeline to fall into the 2.6 to 2.8 coverage range but from weeks 5 to 9 we are generally closing and burning off pipeline [10] at the same rate – hence the coverage ratio is running in a stable, if somewhat tight, range.

Next-Quarter Pipeline
Let’s now look at next-quarter pipeline.  While I think sales needs to be focused on this-quarter pipeline and closing it, marketing needs to be primarily focused on next-quarter pipeline and generating it.  Let’s look at an example:

next qtr pipe

Now we can see that next-quarter plan is $3,250K and we start this quarter with $3,500K in next-quarter pipeline or 1.1x coverage.  The 1.1x is nominally scary but do recall we have 12 weeks to generate more next-quarter pipeline before we want to start next quarter with 3x coverage, or a total pipeline of $9,750K.  Once you start tracking this way and build some history, you’ll know what your company’s requirements are.  In my experience, 1.5x next-quarter coverage in week 3 is tight but works [11].

The primary point here is that given:

  • Your knowledge of history and your pipeline coverage requirements
  • Your marketing plans for the current quarter
  • The trends you’re seeing in the data
  • Normal spillover patterns

That marketing should be able to forecast next quarter’s starting pipeline coverage.  So, pipeline coverage isn’t just an iceberg that marketing thinks we’ll hit or miss.  It’s something can marketing can forecast.  And if you can forecast it, then you adjust your plans accordingly to do something about it.

Let’s stick with our example and make a forecast for next-quarter starting pipeline [12]

  • Note that we are generating about $250K of net next-quarter pipeline per week from weeks 4 to 9.
  • Assume that we are continuing at steady-state the programs generating that pipeline and ergo we can assume that over the next four weeks we’ll generate another $1M.
  • Assume we are doing a big webinar that we think will generate another $750K in next-quarter pipeline.
  • Assume that 35% of the surplus this-quarter pipeline slips to next-quarter [13]

If you do this in a spreadsheet, you get the following.  Note that in this example we are forecasting a shortfall of $93K in starting next-quarter pipeline coverage.  Were we forecasting a significant gap, we might divert marketing money into demand generation in order to close the gap.

fc next qtr

All-Quarters Pipeline
Finally, let’s close with how I think about all-quarters pipeline.

all qtr

While I don’t think it’s the primary pipeline metric, I do think it’s worth tracking for several reasons:

  • So you can see if pipeline is evaporating or sloshing. When a $1M forecast deal is lost, it comes out of both current-quarter and all-quarters pipeline.  When it slips, however, current-quarter goes down by $1M but all-quarters stays the same.  By looking at current-quarter, next-quarter, and all-quarters at the same time in a compact space you can get sense for what is happening overall to your pipeline.  There’s nowhere to hide when you’re looking at all-quarters pipeline.
  • So you can get a sense for the size of opportunities in your pipeline.  Note that if you create opportunities with a placeholder value then there’s not much  purpose in doing this (which is just one reason why I don’t recommend creating opportunities with a placeholder value) [14].
  • So you can get a sense of your salesreps’ capacity. The very first number I look at when a company is missing its numbers is opportunities/rep.  In my experience, a typical rep can handle 8-12 current-quarter and 15-20 all-quarters opportunities [15].  If your reps are carrying only 5 opportunities each, I don’t know how they can make their numbers.  If they’re carrying 50, I think either your definition of opportunity is wrong or you need to transfer some budget from marketing to sales and hire more reps.

The spreadsheet I used in this post is available for download here.

# # #


[1] Assuming you’re in the first few weeks of the quarter, for now.

[2] Which is usually done using forward four quarters.

[3] And ASC follows a normal distribution.

[4] Typically, they are smaller deals, or deals at smaller companies, or upsells to existing customers.  But they’re out there.

[5] Or, whatever your favorite coverage ratio is.  Debating that is not the point of this post.

[6] Once you build up some history you can use coverage ratios to predict sales as a way of triangulating on the forecast.

[7] As a former board member always told me — a quote that rivals “think of salespeople as single-celled organisms driven by their comp plan” in terms of pith.

[8] Or sometimes, fourth-quarter which is another popular pipeline dumping ground.  (As is first-quarter next year for the truly crafty.)

[9] That is, one about how they are going to get their shit together and manage the pipeline better, the first piece of which is getting it clean by week 3, often best accomplished by one or more pipeline scrub meetings in weeks 1 and 2.

[10] Burning off takes one of three forms:  closed/won, lost or no-decision, or slipping to a subsequent quarter.  It’s only really “burned off” from the perspective of the current-quarter in the last case.

[11] This depends massively on your specific business (and sales cycle length) so you really need to build up your own history.

[12] Technically speaking, I’m making a forecast for day-1 pipeline, not week-3 pipeline.  Once you get this down you can use any patterns you want to correct it for week 3, if desired.  In reality, I’d rather uplift from week 3 to get day-1 so I can keep marketing focused on generating pipeline for day-1, even though I know a lot will be burned off before I snapshot my analytics in week 3.

[13] Surplus in the sense that it’s leftover after we use what we need to get to plan.  Such surplus pipeline goes three places:  lost/no-decision, next-quarter, or some future quarter.  I often assume 1/3rd  goes to each as a rule of thumb.

[14] As a matter of principle I don’t think an opportunity should have a value associated with it until a salesrep has socialized a price point with the customer.  (Think:  “you do know it cost about $150K per year to subscribe to this software, right?”)  Perversely, some folks create opportunities in stage 1 with a placeholder value only to later exclude stage 1 opportunities in all pipeline analytics. Doing so gets the same result analytically but is an inferior sales process in my opinion.

[15] Once you’re looking at opportunities/rep, you need to not stop with the average but make a histogram.  An 80-opportunity world where 10 reps have 8 opportunities each is a very different world from one where 2 reps have 30 opportunities each and the other 8 have an average of 2.5.

Ten Ways to Get the Most out of Conferences

I can’t tell you the number of times, as we were tearing down our booth after having had an epic show, that we overheard the guy next door calling back to corporate saying that the show was a “total waste of time” and that the company shouldn’t do it again next year.  Of course, he didn’t say that he:

  • Staffed the booth only during scheduled breaks and went into the hallway to take calls at other times.
  • Sat inside the booth, safely protected from conference attendees by a desk.
  • Spent most of his time looking down at his phone, even during the breaks when attendees were out and about.
  • Didn’t use his pass to attend a single session.
  • Measured the show solely by qualified leads for his territory, discounting company visibility and leads for other territories to zero.

slack boothDoes this actually happen, you think?  Absolutely

All the time.  (And it makes you think twice when you’re on the other end of that phone call – was the show bad or did we execute it poorly?) 

I’m a huge believer in live events and an even bigger believer that you get back what you put into them.  The difference between a great show and a bad show is often, in a word, execution.  In this post, I’ll offer up 10 tips to ensure you get the best out of the conferences you attend.

Ten Ways to Get the Most out of Conferences and Tradeshows

1. Send the right people.  Send folks who can answer questions at the audience’s level or one level above.  Send folks who are impressive.  Send folks who are either naturally extroverts or who can “game face” it for the duration of the show.  Send folks who want to be there either because they’re true believers who want to evangelize the product or because they believe in karma [1].  Send senior people (e.g., founders, C-level) [2] so they can both continue to refine the message and interact with potential customers discussing it.

2. Speak.  Build your baseline credibility in the space by blogging and speaking at lesser conferences.  Then, do your homework on the target event and what the organizers are looking for, and submit a great speaking proposal.  Then push for it to be accepted.  Once it’s accepted, study the audience hard and then give the speech of your life to ensure you get invited back next year.  There’s nothing like being on the program (or possibly even a keynote) to build credibility for you and your company.  And the best part is that speaking a conference is, unlike most everything else, free.

3. If you can afford a booth/stand, get one.  Don’t get fancy here.  Get the cheapest one and then push hard for good placement [3].  While I included a picture of Slack’s Dreamforce booth, which is very fancy for most early-stage startup situations, imagine what Slack could have spent if they wanted to.  For Slack, at Dreamforce, that’s a pretty barebones booth.  (And that’s good — you’re going to get leads and engage with people in your market, not win a design competition.)

4. Stand in front of your booth, not in it.  Expand like an alfresco restaurant onto the sidewalk in spring.  This effectively doubles your booth space.

5. Think guerilla marketing.  What can make the biggest impact at the lowest cost?  I love stickers for this because a clever sticker can get attention and end up on the outside of someone’s laptop generating ongoing visibility.  At Host Analytics, we had great success with many stickers, including this one, which finance people (our audience) simply loved [4].


While I love guerilla marketing, remember my definition:  things that get maximum impact at minimum cost.  Staging fake protests or flying airplanes with banners over the show may impress others in the industry, but they’re both expensive and I don’t think they impress customers who are primarily interested not in vendor politics, but in solving business problems.

6. Work the speakers.  Don’t just work the booth (during and outside of scheduled breaks), go to sessions.  Ask questions that highlight your issues (but not specifically your company).  Talk to speakers after their sessions to tee-up a subsequent follow-up call.  Talk to consultant speakers to try and build partnerships and/or fish to referrals.  Perhaps try to convince the speakers to include parts of your message into their speech [5].

7. Avoid “Free Beer Here” Stunts.  If you give away free beer in your booth you’ll get a huge list of leads from the show.  However, this is dumb marketing because you not only buy free beer for lots of unqualified people but worse yet generate a giant haystack of leads that you need to dig through to find the qualified ones — so you end up paying twice for your mistake.  While it’s tempting to want to leave the show with the most card swipes, always remember you’re there to generate visibility, have great conversations, and leave with the most qualified leads — not, not, not the longest list of names.

8. Host a Birds of a Feather (BoF).  Many conferences use BoFs (or equivalents) as a way for people with common interests to meet informally.  Set up via either an online or old-fashioned cork message board, anyone can organize a BoF by posting a note that says “Attention:  All People Interested in Deploying Kubernetes at Large Scale — Let’s Meet in Room 27 at 3PM.”  If your conference doesn’t have BoFs either ask the organizers to start them, or call a BoF anyway if they have any general messaging facility.

9. Everybody works. If you’re big enough to have an events person or contractor, make sure you define their role properly.  They don’t just set up the booth and go back to their room all day.  Everybody works.  If your events person self-limits him/herself by saying “I don’t do content,” then I’d suggest finding another events person.

10.  No whining.  Whenever two anglers pass along a river and one says “how’s the fishing?” the universal response is “good.”  Not so good that they’re going to ask where you’ve been fishing, and not so bad that they’re going to ask what you’ve been using.  Just good.  Be the same way with conferences.  If asked, how it’s going, say “good.”  Ban all discussion and/or whining about the conference until after the conference.  If it’s not going well, whining about isn’t going to help.  If it is going well, you should be out executing, not talking about how great the conference is.  From curtain-up until curtain-down all you should care about is execution.  Once the curtain’s down, then you can debrief — and do so more intelligently having complete information.


[1] In the sense that, “if I spend time developing leads that might land in other reps’ territories today, that what goes around comes around tomorrow.”

[2] In order to avoid title intimidation or questions about “why is your CEO working the booth” you can have a technical cofounder say “I’m one of the architects of the system” or your CEO say “I’m on the leadership team.”

[3] Build a relationship with the organizers.  Do favors for them and help them if they need you.  Politely ask if anyone has moved, upgraded, or canceled their space.

[4] Again note where execution matters — if the Host Analytics logo were much larger on the sticker, I doubt it would have been so successful.  It’s the sticker’s payload, so the logo has to be there.  Too small and it’s illegible, but too big and no one puts the sticker on their laptop because it feels like a vendor ad and not a clever sticker.

[5] Not in the sense of a free ad, but as genuine content.  Imagine you work at Splunk back in the day and a speaker just gave a talk on using log files for debugging.  Wouldn’t it be great if you could convince her next time to say, “and while there is clearly a lot of value in using log files for debugging, I should mention there is also a potential goldmine of information in log files for general analytics that basically no one is exploiting, and that certain startups, like Splunk, are starting to explore that new and exciting use case.”

Using Time-Based Close Rates to Align Marketing Budgets with Sales Targets

This post builds on my prior post, Win Rates, Close Rates, and Milestone vs. Flow Analysis.  In it, I will take the ideas in that post, expand on them a bit, and then apply them to difficult problem of ensuring you have enough marketing demand generation budget to hit your sales targets.

Let’s pretend it’s 4Q17 and that we need to model 2018 sales based solely on marketing-generated SALs (sales accepted leads).  To do that, we need to decompose our close rate over time because knowing we eventually close 40% of SALs is less useful than knowing the typical timing in how they close over time.

decompose closed

In a perfect world, we’d have 6-8 cohorts, not two.  The goal is to produce the last line, the average of the in-quarter, first-quarter, second-quarter, and so on close rates for a SAL.

Using these time-based average close rates, we can build a waterfall that takes historical, forecast (for the current quarter), and planned 2018 SALs and converts them into deals.


This analysis suggests that with the currently planned SALs you can support an ARR number of $16.35M.  If sales needs more than that, you either need to assume an improvement in close rates or an increase in SAL generation.

Once you’ve established the required number of SALs, you can then back into a total demand-generation budget by knowing your cost/SAL, and then building out a marketing mix of programs (each with their own cost/SAL) that generates the requisite SALs at the targeted overall cost.