Using automation to help with bid / no-bid decisions – Can this work?

Fergal McGovern

CEO & Founder

5 min read
Automation - Bid no bid decisions

Blog UPDATE: We ran a webinar on April 10th where Jim Creutz (subject of this post) showed what he is doing. You can get the presentation slides here to see if using automation to help a bid manager with bid / no-bid decisions.

In early 2011, FIT systems, a Florida-based SDVOSB (service-disabled veteran-owned small business) asked these questions. Can we:

  1. Quickly determine whether to bid / no-bid or team
  2. Save time and avoid reading poorly qualified solicitations.
  3. Compare multiple opportunities for a ‘good fit’ in a few minutes.

In February 2012, Jim Creutz, Business Development Director at FIT Systems and I got chatting. Jim shared what they were doing with VisibleThread to support these efforts. I was intrigued.

Of course, they were using VT for the normal stuff; making better quality props with readability analysis, generating starter compliance matrices in minutes, and driving more efficient pink and red team reviews. But, Jim showed me how they were also using VisibleThread to augment the very early stages of Capture Planning too to help bid / no-bid and teaming decisions. Now, this was new!

They had started using language analysis to scan solicitation docs for what Jim called ‘affinity’. Basically, he was identifying the right opportunities to bid or team on with language indicators. As I blogged here on win themes, it is always amazing to see how customers apply VisibleThread, so this was another really interesting case.

So, if you are a Capture Manager or Proposal Manager involved in Business Development at all, you will find this post useful. If you want to see this technique in the flesh, Jim shared what he is doing in a webinar we ran on the 10th of April. Click here for the slides.

I am very interested to hear whether you have attempted to analyze solicitation docs using a similar approach. If so, how did it go? Let us know your thoughts on this approach. Do you think it can work in your case? Is it helpful? Would it save time? Post comments and let us know.

So, let’s get back to Jim. Here’s how he did it.

Step 1: Define Affinity Characteristics & weightings

Jim set up what he calls affinity characteristics in a spreadsheet.

Here are three categories he defined:

  • Experience
  • Technical
  • Business Attributes

Under each category, he listed phrases and terms that would indicate ‘fit’ or ‘lack of fit’. Here is a sample of some of the characteristics he defined:

TechnicalBusiness AttributesExperience
JavaDUNSAustin Automation Center
JAVA EECAGEBenefits Delivery Network
REQUIREMENTS IDENTIFICATIONFlorida Management Services ContractC&P
Mercury InteractiveMinority BusinessCompensation Pension
metrics developmentSBAFASS
object modelsSEAPORTeORM
OLTPService Disabled Veteran OwnedShared Corporate Database
Transaction ProcessingSDVOSBVAA tools
knowledge management541511VBA Corporate
IT services541513VBA Applications Architecture
Oracle Data Base541519VETSNET
Schedulingdisadvantaged businessVR&E
skill transferFacility ClearanceVeterans Benefits Administration
Sun SolarisFFP 
SDLCSubcontract management 
test datatop Secret 

In total, Jim defined a total of 77 technical indicators to scan for, 23 individual business indicators, and 15 experienced indicators.

Next, for each indicator, Jim assigned a weighting. Positive weightings between 1 and 5 indicate a good fit, with 5 as the best fit.

Negative weightings indicate poor fit. These ranged from -1 to -5, with -5 as the worst fit. So, now Jim had an Excel spreadsheet with categorized indicators and weightings.

Jim also defined a categorized ‘Concept List’ in VisibleThread so that he would be able to scan for the indicators. The concept list will allow us to explicitly scan the documents later as part of step 3. So, onto the next step, uploading docs…

Step 2: Take your docs from GovWinIQ or FedBizOpps and upload them to VisibleThread

Jim’s company subscribes to GovWinIQ from Deltek. This means they have a steady stream of solicitations coming in on a daily basis. For those not using Deltek’s product, you can still get your solicitations from FedBizOpps of course.

So, Jim takes the solicitation docs that are worth a second look and uploads them to a VisibleThread folder.

He tends to name his folders either by date range or, if a large number are coming at him from a specific agency/dept., he’ll add the agency name. So he might have folders called ‘2012 Week April 2  – Solicitations’ or ‘Army Solicitations – April 2012’ or ‘Navy Solicitations – April 2012’ etc.

By adding to categorized folders like this, over time he can assess trends also. But that’s a separate benefit.

Step 3: Run a Scan

Next Jim runs the scans against the doc. You might recall back in step 1, Jim had defined a ‘concept list’ holding his technical, business and experience indicators. So, a scan is really a way to uncover any references to those indicators.

So, scanning is pretty simple. He does this:

  1. He selects the folder. He clicks ‘Run Concept List Analysis…’, selects his ‘affinity’ concept list, and then.
  2. He reviews Concept Coverage across the uploaded docs. In other words, he is presented with a matrix-type view.

The screenshot below is an example of how this looks in VisibleThread. Listed across the top are the documents; RFPs, SOWs, and PWSs for candidate solicitations.

The left shows the indicators; technical business and experience. Each column represents a solicitation document. The numbers indicate coverage (number of hits).

bid decisions

Step 4: Crunch the numbers in Excel

Finally, Jim now takes the above information and exports it to Excel.

He places it in the context of a detailed spreadsheet he has developed.

Jim’s final output is what he calls a ‘normalized score’. This is a calculated number based on the sum of each attribute’s Frequency Density X Weighting.

Now, this may sound a little tricky, but here’s how he did it:

  1. Take the weighting as outlined earlier, 1 to 5 for positive, -1 to -5 for negative.
  2. Calculate Density by taking the total word count in the document and devising a % based on (number hits / total word count *100). This gives him a density percentage value.
  3. Multiple the weighting against the density percentage value to give a normalized score for each attribute.

Here’s an example Excel report for one of Jim’s solicitations:

bid decisions - Jim Creutz's Solications

Jim has also developed a bunch of excel sheets that can show how all the solicitations compare graphically.

Here we see 6 solicitations compared by ‘Business Attributes’ using one of Jim’s comparisons. The green checks should be considered further. The ‘question’ one may be worth a read. We should likely pass on the red x solicitations.

Comparing solications by business attributes - VisibleThread

If you want to learn more about this, Jim shared what he is doing in a webinar we ran on the 10th of April

What has changed for Jim and his team?

The business benefits for Jim are very clear:

  1. Time savings – Every day one of Jim’s colleagues had been spending between 1 and 2 hours reviewing new opportunities. Now that time is cut, because 1.) they are reading fewer more qualified solicitations and 2.) it is faster to identify exactly where the hot buttons are.
  2. Improve Win Rate – Better decisions around the right bids to pursue – using objective metrics for ‘apples to apples’ comparisons of multiple opportunities.
  3. Pass Fast on inappropriate stuff – Faster turnaround time in determining when to dig deeper and when to pass. If you know when to pass it allows you to focus on better-qualified opportunities.
  4. Better win theme development – This is a side effect of Jim’s work. Because he’s got a more detailed sense of what the hot buttons in the solicitations are, he can now do a much more accurate job of establishing win themes and items to emphasize in the response.

Summary Takeaways

I will leave you with some of my own summary observations:

  • Jim has shown how automation can drive very quick qualification cycles. While his math skills are to be applauded, any proposal shop could apply this with VisibleThread. Drop us a note if you want help.
  • Jim has evolved one way to do the task that is performing well ‘under fire’, and in very timeframes of under 5-7 minutes for collections of quite large docs.
  • The business benefits of chasing the right opportunities are profound.

So, to see it in action, see Jim’s webinar slides and recording here.


Book a Demo