Finding companies by the software they use is one of the highest-intent ways to build a B2B list. You are not asking "who is in this industry?" You are asking "who already runs the system my product connects to, replaces, or sits beside?"
Start with one tool, not a category soup
Pick Salesforce, Shopify, Cloudbeds, Epic, or whatever your product actually cares about. Search that tool first in technologies, then add industry and size in company search.
"Uses cloud software" is not a filter. It is a shrug.
Add the market so the list can be worked
Technology-only lists explode. Country and employee band bring them back to a number a human can finish. If you sell in the UK mid-market, say so in the filters before you reveal.
Use missing software as a filter too
The best lists are often "has X, does not have Y." Salesforce without a MAP. Hotel on OTAs without a booking engine. Shopify without a retention tool. That pattern is a story. Presence alone is only a starting point.
Read the profile before you scale
Confirm the company is still in the industry you think it is. Confirm the website matches. Confirm the tool still looks current. Then export. If you need this done as a file, use custom data services.
Write outreach that could not be swapped to another account
If you can paste the first line onto a different company in a different stack, rewrite it. The software they use is the detail that earns the rest of the email.
A 30-minute version of this workflow
Tool, industry, size, geography, sample reveal, ten personalized notes. That is enough to learn if the hypothesis is real. There is a longer walkthrough in how to build a target account list.
What this method is not
It is not a guarantee they will buy. It is not a replacement for talking to people. It is a way to stop starting from a blank industry list when your product only makes sense in a certain environment.
Building an exclusion list that protects your domain
Every technology-led list should include who you will not contact: current customers, open opportunities, companies that already own your category, and industries you cannot support. Put those exclusions in the search or in the export review.
Finding companies by the software they use is powerful enough to create a lot of bad sends quickly. The filter that saves you is the one that removes the obvious no.
Save that search. Name it after the offer, not after the tool. "Salesforce cleanup - UK mid-market" is a campaign. "Salesforce users" is a dump.
From software list to weekly habit
Monday: refresh the saved search. Tuesday: reveal a small batch. Wednesday: send. Thursday: follow up. Friday: mark what the stack got right. That habit beats a quarterly data project.
If the habit works, expand. If it does not, change the tool or the offer, not the volume. Software-led prospecting fails when people scale a weak first line.
Worked example: one tool, two offers
Take Shopify. Offer A is for stores that look thin: payments and email are light. Offer B is for stores that already look like a stack and need analytics or ops. Same software. Two lists. Two first lines.
Finding companies by the software they use only works if you split offers like that. Otherwise you send a sophisticated pitch to a shop that needs a basic integration, and a basic pitch to a team that will ignore you.
Build both searches. Run the one you can support this month. Park the other. Software-led prospecting is a queue, not a bonfire.
Implementation notes for software-led prospecting
Create a library of saved searches named by offer. If the names are tool names only, people will merge them. Finding companies by the software they use is an offer motion that happens to start with a tool.
Add a monthly audit: open ten random revealed accounts and ask if the software still looks current. If three are wrong, your messaging is lying at a 30 percent rate. Fix detections or fix copy.
Never let a new hire send from a raw technology export on day one. They should watch someone read a profile first. The method is simple. The judgment is the product.
Pair each software list with one exclusion list. Presence without absence is how you email people who already bought the category.
Report pipeline by software pattern. Leadership understands "Shopify thin-stack" faster than they understand "technographic segment 4."
When the same search is rebuilt every week, automate it with a saved view or a custom list. Repetition is a sign the hypothesis worked. Treat it like production, not like research.
Field manual for software-led weeks
One tool per week until a tool produces meetings. Then stay on that tool for a month. Finding companies by the software they use is a focus sport.
Write the exclusion list before the inclusion list. It is the difference between a prospecting motion and a spam incident.
Read ten profiles out loud with another person. If you cannot say why each company is on the list, remove it.
When a detection looks stale, skip the account. One wrong stack mention trains the market to ignore you.
Promote winning searches to saved views or custom lists. Retire losing searches. The library should stay small on purpose.
Twelve-step software-led playbook
This is how to find companies by the software they use without creating a spam incident. The software is the start. The exclusions and the first line are the craft.
Teach the playbook as a weekly ritual, not as a one-off project. Software changes. Job ads change. Your saved search should be re-read, not blindly re-exported.
If two tools seem equally good, run two weeks, not one merged week. Merged weeks hide which software pattern created the meeting.
When leadership asks for a bigger technology universe, show them the scoreboard for the current tool. Expand after a win. Expanding to look busy is how technographic search gets a bad name.
The last rule: if you cannot name the software and the missing layer in the first line, you are not ready to send. Go back to step two and write the offer again.
- Pick one tool your product actually cares about.
- Write the offer that tool implies, including the missing layer if there is one.
- Search the tool, then add one industry and one size band.
- Add geography so a human can finish the list.
- Write exclusions before you reveal.
- Browse and drop rows you cannot explain.
- Reveal a small batch and visit the sites.
- Send notes that name the software and the job.
- Log stale detections and skip those accounts.
- Keep one tool per week until a tool produces meetings.
- Save the winning search under the offer name.
- Promote repetition to a custom list so you stop rebuilding it.
Presence, absence, and why both filters matter
How to find companies by the software they use is only half the search. The other half is finding companies that do not use the layer you sell. Presence without absence is how you email people who already bought the category.
A useful cut sounds like: uses Salesforce, does not show a marketing automation platform, sits in a size band you can implement. Or: uses a hotel PMS, does not show a booking engine. The software they use is the door. The software they lack is the reason to walk in.
Build the inclusion in technographic filters, then add industry and size in company search. If the result count is huge, add geography before you reveal. A software-led list that a human cannot finish will be sprayed.
Write exclusions for customers, open opportunities, and category owners before the first export. Finding companies by the software they use without those exclusions is a spam incident waiting for a send date.
One tool, two offers, two lists
The same software can imply two different jobs. A thin Shopify store may need payments or email. A full Shopify stack may need analytics or operations. If you keep those accounts in one export, the first line becomes mush and the meeting rate falls.
Split the lists. Run the offer you can support this month. Park the other. Software-led prospecting is a queue. Teams that treat it like a bonfire burn the domain and then blame the detection.
Name saved searches after the offer, not only after the tool. "Shopify thin-stack email" can be measured. "Shopify users" cannot. When leadership asks for a bigger technology universe, show the scoreboard for the current offer first.
Keeping software-led lists from going stale
Tools get ripped out. Job ads linger. Footer scripts linger. Once a month, open ten random revealed accounts and ask if the software still looks current. If three are wrong, your messaging is lying at a thirty percent rate.
Skip stale-looking rows. One wrong stack mention trains the market to ignore you. Update the saved search the same day you learn an exclusion. Finding companies by the software they use is a weekly ritual, not a one-off project.
When the same search is rebuilt every week, promote it to a saved view or a custom company list. Repetition means the hypothesis worked. Treat it like production.
PrimoDato Editorial Team
B2B Intelligence & Prospecting Research
The PrimoDato team writes about company data, B2B prospecting, technographic intelligence, and sales strategy based on what we see across our platform and the markets we cover.