iHousedesign · Strategic Canvas

Fashion Retail
Product Ideas

Software products and productized services for fashion boutiques and DTC brands. Ranked by asset leverage, distribution viability, and time to revenue.

Date May 13, 2026
Research 107-source corpus
Status Pre-validation
Tier 1 — Primary Candidates
Intelligence Subscription · Recurring
Practitioner Reality Index
72–78 / 100
Monthly intelligence subscription distilling tactical reality from the fashion retail corpus. Each issue answers one specific question operators actually ask — what return rate is normal at a given AOV, which workarounds the median performer runs, whether brands at $20M actually profit on first purchase. Positioned explicitly against vendor analytics, which the corpus shows operators universally distrust. Direct analog to BaseRate Reality Checker for M&A — same structure, warmer distribution.
corpus leverage editorial fit no platform needed warm distribution needs subscriber base
Pricing
$299–599/mo (Brand Lens: $1,500–2,500/mo)
Target MRR
50–80 subscribers → $25–40k
Assets fired
Corpus · pipeline · editorial design · fashion network
Kill criteria
Fewer than 5 paying subscribers in 30 days post-launch
Productized DFY Service · Recurring
Cinderella Campaign Service
74–79 / 100
Done-for-you monthly email service for Shopify fashion brands. Client provides monthly CSV export from POS + Shopify (3 clicks, familiar motion). Agency builds size-matched segments and runs 4–6 targeted campaigns liquidating broken size runs at full margin. Example: 24,000-person list → 148 exact-fit buyers for remaining Size 10 inventory. Research-validated pain: forced markdowns destroy margins from 60% → 35–40%. No POS integration needed — CSV works fine. No platform dependency.
fastest revenue research validated Adaveo framework ready Shopify API available requires campaign execution
Pricing
$2,500–4,000/month retainer
Time to revenue
2–3 weeks (manual-first delivery)
Data access
Monthly CSV export — no POS integration required
Target clients
Shopify fashion brands $5–30M revenue
Tier 2 — Strong Secondary
Productized Service · One-time + Renewal
Returns Reduction Audit
68–74 / 100
One-time audit of 12 months of return data. Identifies top 5 SKUs generating 80% of returns, diagnoses root cause (sizing inconsistency, photo misrepresentation, description gap), delivers 30-page PDF + 60-min call. Research validated: 16% non-food e-commerce return rate, brands "underwater on reverse logistics," would be better off discarding returned goods. No incumbent audits the diagnosis layer — Loop and Returnly handle the transaction, not the intelligence.
no platform needed high-urgency pain manual-first viable proof of playbook needed first
Pricing
$4–7k one-time · $12–20k/year with quarterly renewal
Kill criteria
No measurable improvement in 90 days post-audit
Intelligence Subscription · Quarterly
Workaround Briefing
55–62 / 100
Quarterly intelligence briefing tracking which platform workarounds are working (Meta dummy ad accounts, branded search isolation, dark posting, mystery boxes for dead stock), which have become liabilities, and which carry new platform risk. Research shows 90% of brands rely on undocumented workarounds. Niche but real. Better as an upsell to the Practitioner Reality Index than as standalone.
unique angle narrow buyer upsell only
Pricing
$800–1,500/quarter
Position
Add-on to Practitioner Reality Index, not standalone
Tier 1 — Primary Candidates (continued)
AI Visual Content · Productized Service · Recurring
AI Campaign Imagery Studio
70–76 / 100
Monthly productized service producing AI-generated campaign imagery for fashion brands: product placement on AI influencer models, still images formatted for email newsletters and Instagram, and short video content for Reels and TikTok. Uses trained models to place client products onto AI-generated figures with brand-consistent aesthetics. Directly answers the corpus-validated content treadmill pain — founders explicitly say they did not sign up to be content creators, the volume expectation is described as "wild," and traditional photography costs and turnaround times push brands toward cheap lofi workarounds. This replaces the workaround with a professional, affordable, repeatable alternative.
editorial credibility (Vogue Russia) AI model capability in-house corpus-validated pain natural upsell from Cinderella recurring monthly must stay productized — not custom client-work drift risk
What it produces each month
Fixed monthly deliverable — not open-ended creative services. Three package tiers, all with defined scope: Tier 1 delivers 12 static images (6 email-formatted, 6 Instagram-formatted). Tier 2 adds 4 short Reels/TikTok clips (15–30 seconds, product-focused). Tier 3 adds a branded AI influencer persona trained on the client's aesthetic, used consistently across all assets. Client provides: product photos or 3D renders, brand color palette, one-paragraph brief. You produce everything else. No discovery calls, no revision rounds beyond one round per delivery cycle.
Why AI influencers specifically
Human influencer partnerships for fashion brands at the $5–30M revenue tier are expensive, slow to negotiate, and unpredictable. The talent owns their image and audience. AI influencers solve all three problems: the brand owns the model entirely, shoots happen in hours not weeks, and the aesthetic is fully controllable. The corpus validates that brands are already running lofi workarounds (fake iMessage overlays on static images, reviving dead ads, dark posting through third-party handles) precisely because production capacity is the bottleneck. AI imagery removes that bottleneck at a fraction of traditional photography cost.
The critical constraint: productized, not custom
The only scenario where this slides back into client work is if you say yes to custom requests outside the package scope — a different model aesthetic, a different format, a one-off campaign concept. The defense is a strict package menu with explicit out-of-scope pricing (e.g., custom AI model training is a one-time $2,000 setup fee, billed separately). Everything within the package is delivered without creative negotiation. Clients who need total creative control hire a photographer. Clients who need reliable, beautiful, fast, affordable content subscribe to the package.
Pricing (USD)
Tier 1: $800/mo · Tier 2: $1,500/mo · Tier 3: $2,500/mo
Target MRR
10 clients avg $1,500 = $15k MRR. Stacks with Cinderella clients naturally.
Asset leverage
Vogue Russia editorial eye · AI model training capability · fashion brand taste · design skill
Acquisition
Same LinkedIn outreach list. Cinderella clients become natural first buyers — campaigns need imagery.
Connection to other products
Cinderella clients need email imagery monthly. Returns Audit fixes photo misrepresentation → this produces the fix. PRI subscribers see content cost data → hire you.
Kill criteria
Fewer than 3 paying clients in 60 days, or scope creep consuming more than 40% of delivery time. If custom requests dominate, raise out-of-scope pricing until they stop.
Warning: this is the product most likely to drift into client work. Fixed packages and hard scope boundaries are not optional — they are the entire structural defense against recreating the photography agency you left behind.
Killed — Do Not Build
Software Platform
Adaveo CDP / Unified Inventory Platform
Not viable now
Full software platform connecting POS + e-commerce + CRM for fashion retailers. Requires POS integration (Adel/Philippe dependency unresolved, no confirmed NA integrations), multi-year build, and behavior change from overwhelmed operators. Incumbents well-funded. The underlying pains are real but the product form is wrong — same outcomes achievable via the Cinderella Campaign service without the infrastructure trap.
Kill reason: infrastructure-first trap · POS dependency unresolved · commodity positioning · platform distrust documented in corpus
Software Tool
AI Shopping Agent / Chatbot for Retail
Not viable
The corpus explicitly flags this category as pre-poisoned. Operators and consumers distrust AI agents — hallucinations, "corruption of recommendations," fear of paid prioritization over genuine fit. Building into active distrust is structural disadvantage regardless of quality.
Kill reason: category distrusted by corpus — operators named it specifically
Core Principles from Research
Recommended path

Primary: Practitioner Reality Index — BaseRate model applied to fashion with warmer distribution. Build the corpus, launch landing page, target 5 paying subscribers in 30 days via fashion network.

Cash flow secondary: Cinderella Campaign service — fastest path to first revenue, validates willingness to pay for outcomes before building anything. Run both in parallel only if distribution is separate. Do not split execution attention between them.

Customer Acquisition — How to Get There
Practitioner Reality Index · Acquisition Path
PRI: Zero to 25–40k MRR
80–100 subscribers needed
At $299–599/month average, you need 50–80 paying subscribers for $25k MRR. At $1,500–2,500/month Brand Lens tier, 15–20 clients gets you there. The math favors hunting for Brand Lens clients early rather than volume of cheap subscribers.
Phase 1 — First 5 Paying Subscribers (Days 1–30)
Write Issue 1 in full before doing any outreach. This is non-negotiable — you cannot sell something that doesn't exist yet. Send it cold to 40–50 qualified fashion brand operators via LinkedIn with one line: "I'd like your honest feedback on this — 8 minutes to read." No pitch, no pricing, no ask. Watch who reads it. Follow up only with those who engage. Ask one question: "Would you pay CA$300–600/month for this monthly?" First five who say yes become founding members at a locked lower rate. This is your proof of concept.
Phase 2 — Subscribers 5 to 25 (Days 30–90)
LinkedIn organic content becomes the engine. Post one corpus finding per week publicly — not the full issue, just the sharpest single insight. Format: one claim, one number, one implication. "107 operator interviews. Only 3 of 8 standard metrics are actually trusted by the people running these brands. Here's which 3." This builds an audience of exactly the people who would subscribe. Direct them to Substack. Every post is a free sample. No ads at this stage — organic only, zero cost of acquisition until you have proof the content converts.
Phase 3 — Subscribers 25 to 80 (Days 90–180)
Substack becomes the home. Free tier delivers one teaser insight per month. Paid tier ($299–599) delivers the full issue. Substack's internal discovery and recommendation engine starts working once you have 20+ paid subscribers — it actively surfaces your publication to relevant readers. Add a Brand Lens tier ($1,500–2,500/month) for brands that want their specific revenue tier or category benchmarked directly. This tier requires a 30-minute intake call but converts at high value. Three Brand Lens clients alone = $4,500–7,500 MRR.
Ads?
No. Not until month 4+. CAC via ads for B2B intelligence is $200–800+. Organic LinkedIn + Substack costs zero until proven.
Cost per subscriber (organic)
CA$0 direct cost. Time cost: ~3 hrs/week for content + outreach. LTV at $400/mo avg: CA$4,800/year.
TikTok / Instagram?
Skip. Wrong format for $299–599/mo B2B product. Fashion operators use both but not to find business subscriptions.
Telegram?
Delivery channel only for paid subscribers — private group as a bonus perk. Not an acquisition channel.
Blog / SEO?
Long game only. SEO takes 6–12 months to compound. Useful in Phase 3+ but irrelevant for first 90 days.
Kill criteria
Fewer than 5 paying subscribers after sending Issue 1 to 50 prospects. If zero interest from warm outreach, the format is wrong — not the channel.
Cinderella Campaign Service · Acquisition Path
Cinderella: Zero to First 8 Clients
8 clients = $20–32k MRR
This is a high-ticket service ($2,500–4,000/month), not a subscription product. You don't need volume — you need 8 clients. The acquisition model is completely different from PRI. No content engine needed. No Substack. Pure direct outreach.
Phase 1 — First Client (Days 1–21)
Build one screenshot showing the Cinderella logic visually: a bar showing 24,000 total list → filtered to 148 exact-fit buyers for a specific remaining inventory item. No copy, no long explanation. One image that makes the math obvious. Send it via LinkedIn to 50 Shopify fashion brand founders at $5–30M revenue with a single line: "We built this for a client last month. Their remaining Size 10 inventory sold at full margin in 4 days. Worth a 20-minute call?" That is the entire pitch. The image does the work.
Phase 2 — Clients 2 to 8 (Days 21–90)
First client becomes your case study immediately. "Client cleared CA$14,000 of stuck inventory in 4 days at full margin — no markdown." That number goes into every subsequent outreach message. Each closed client funds the next month of outreach. No ads needed. Referrals from first 2–3 clients carry disproportionate weight in the fashion boutique community — it's a small, interconnected world where word travels fast if results are real.
Ads?
No. A $3,000/month service sells person-to-person, not via Meta campaigns. Ads are for sub-$300 products.
Cost per client (organic)
CA$0 direct. Time: ~5 hrs outreach to land one client. LTV at $3,000/mo × 12 months: CA$36,000.
Outreach volume needed
50 contacts → ~5 replies → ~2 calls → 1 close. Repeat monthly until 8 clients.
Primary channel
LinkedIn direct message with visual proof. Secondary: warm referrals from first 2 clients.
Crossover with PRI
Cinderella clients become natural PRI subscribers. Same buyer, different budget line. Bundle: service client gets PRI subscription included.
Kill criteria
Zero demo calls from 150 outreach contacts over 3 weeks. If no one books a call after seeing the visual proof, the offer framing is wrong.
Returns Reduction Audit · Acquisition Path
Returns Audit: Zero to First 6 Clients
6 audits = $24–42k / year
One-time $4–7k engagement, not a monthly retainer. Different acquisition motion than the other two products. You are selling a diagnosis, not a service. The buyer is a founder who already knows their returns are destroying margin but doesn't know exactly why or which SKUs are responsible. That person exists at every Shopify fashion brand doing $5M+ in revenue.
Phase 1 — First 3 Clients (Days 1–45)
The outreach asset here is a single provocative number pulled from your corpus: "The average fashion e-commerce brand loses more money processing returns than it recovers from reselling them. We audit the 5 SKUs responsible for 80% of the damage." Send this via LinkedIn to the same founder/operator list used for Cinderella. No visual mockup needed — the stat does the work. First 3 audits should be priced at a discount ($2,500–3,000) to build the playbook. You do not yet know exactly what the deliverable looks like until you run it twice. Do not charge full price before you have a repeatable process.
Phase 2 — Clients 4 to 6 and Renewal Logic (Days 45–120)
After three audits you have a case study with a real number: "Client reduced return rate from 22% to 13% on their top-selling outerwear SKU within 60 days of implementing audit recommendations." That stat goes into all subsequent outreach. Pricing moves to full rate ($4–7k). Quarterly renewal framing: "We recommend re-running the audit every 90 days as new SKUs enter the catalog." This converts a one-time engagement into $12–20k/year. Returns Audit clients naturally become Cinderella Campaign clients — they've already shared their inventory and customer data with you.
Ads?
No. Same logic as Cinderella — high-ticket one-time sale needs a conversation, not a click.
Cost per client (organic)
CA$0 direct. First 3 at reduced price build the playbook. Full-rate from client 4 onward.
Outreach volume needed
Same list as Cinderella. 50 contacts → 4–6 replies → 2–3 calls → 1 close at this price point.
Crossover with other products
Returns Audit → Cinderella Campaign (data already shared). Both → PRI subscription. Natural upsell ladder.
Kill criteria
First 3 audits don't produce measurable return-rate improvement within 90 days. If the playbook doesn't work, don't sell more.
Sequencing note
Do not launch Returns Audit before Cinderella is running. Same buyer, same data, but Cinderella validates the relationship first.
Delivery & Channel Architecture — All Three Products
Practitioner Reality Index · Delivery
PRI: How Content Reaches Subscribers
PRI is the only product that requires a delivery infrastructure. The other two are service engagements delivered over calls and PDF documents. PRI is an ongoing publication.
Primary delivery: Email via Substack
Every issue is delivered to paid subscribers as a long-form email. Substack handles the payment processing, subscriber management, and email delivery in one place. No separate email platform (Klaviyo, Mailchimp) needed at this stage — that would be infrastructure overhead for a product that hasn't proven itself yet. Substack also has an internal discovery algorithm that surfaces your publication to readers who follow similar newsletters. Once you cross 20–30 paid subscribers it starts recommending you automatically to relevant readers. This is free compounding distribution you don't have to manage.
Why not Telegram for delivery?
Telegram works well for crypto signals, trading alerts, and communities where real-time push notifications add value. Fashion retail operators are not checking Telegram for business intelligence — they read their inbox in the morning. Telegram also creates a group-chat dynamic where your content gets buried in discussion threads, diluting the perceived value of the publication. The one legitimate use of Telegram: a private channel (broadcast only, no replies) for paid subscribers as a bonus delivery layer — they get a push notification when a new issue lands. This is a perk, not the primary delivery mechanism. Never the primary.
Why not WhatsApp group?
WhatsApp groups create an obligation to respond. The moment you create a WhatsApp group for subscribers, they expect you to be present in it, answer questions, and moderate discussion. That is a community product, not an intelligence publication — completely different model with completely different operational demands. WhatsApp also has no paywall capability, no subscriber management, and no analytics. It is the wrong infrastructure for a paid intelligence product at any price point.
Why not a Facebook group?
Facebook groups are used by fashion boutique owners — the corpus actually mentions them (the "Secret Closet" clearance groups). But they're used for operations and peer support, not for paying for intelligence. A Facebook group positions you as a community moderator, not an authority. It also makes your content free the moment it's posted in the group. The right use of Facebook: join existing fashion boutique owner groups as a participant, share one corpus finding occasionally, and direct interested people to your Substack. You are a guest in those groups, not the host.
Secondary distribution: LinkedIn organic posts
Once per week, post one finding from the corpus publicly on LinkedIn. Not the full issue — the single sharpest number or counterintuitive insight. This functions as a free sample that drives Substack subscriptions. The LinkedIn post is the top of the funnel. The Substack is the conversion point. No other platform is needed until month 4+.
What about TikTok and Instagram?
These are correct platforms for reaching fashion brand operators — they all use both. But they are not where a $299–599/month B2B intelligence product gets discovered. Fashion operators scroll TikTok for creative inspiration and competitor observation, not to find intelligence subscriptions. Instagram is the same. If you post corpus insights on these platforms you may build a following of fashion enthusiasts, fashion students, and small boutique hobbyists — none of whom will pay $300/month. You would be building the wrong audience at significant time cost. Deferred to month 6+ as brand-building only, not acquisition.
What about a blog and SEO?
SEO is a 6–12 month compounding strategy. It costs nothing but time and produces nothing for the first 90 days. Do not prioritize it before you have paying subscribers. The correct sequence: prove someone pays → then build SEO content to attract more of those people. Building SEO content before proof is the content version of the infrastructure trap. When you do start (month 4+), the corpus gives you a massive advantage — you have 107 sources of practitioner data to write from, which no competitor can replicate.
Cinderella Campaign + Returns Audit · Delivery
Services: How Work Gets Delivered to Clients
Both Cinderella and Returns Audit are service engagements. They have no delivery infrastructure requirements — no newsletter platform, no community, no content engine. Delivery is direct and personal.
Cinderella delivery
Client sends monthly CSV export (Shopify + POS, 3 clicks). You build segments, write 4–6 email campaigns, and either send them via the client's existing Klaviyo account or set up a new one. Results report delivered as a one-page PDF: campaigns sent, open rates, revenue attributed, inventory cleared. Monthly. That is the entire delivery. No portal, no dashboard, no app — until you have 10+ clients and the reporting becomes genuinely repetitive.
Returns Audit delivery
Client sends 12 months of return data (CSV export from their returns platform — Loop, Returnly, or manual Shopify export). You analyze, produce a 20–30 page PDF: top 5 problem SKUs, root cause per SKU (sizing inconsistency vs. photo misrepresentation vs. description gap), and a concrete remediation checklist. Delivered with a 60-minute call to walk through findings. That is the full engagement. No recurring delivery until quarterly renewal.
Why no Telegram, WhatsApp, or Facebook group for services?
Service clients communicate over email and scheduled calls. A WhatsApp group for service clients creates always-on expectations and blurs professional boundaries. Telegram adds no value over email for asynchronous document delivery. The right communication stack for services: email for async, Zoom or Google Meet for calls, Loom for async video walkthroughs of deliverables. Simple. Nothing else.
Channel verdict — all three products

Ads: no for all three. Wrong economics at every stage before 20+ paying clients or subscribers. Organic LinkedIn costs zero and reaches the exact decision-maker you need.

PRI: LinkedIn outreach (acquisition) → Substack (home base, payments, email delivery) → LinkedIn organic posts weekly (top of funnel). Telegram as optional push-notification bonus for paid subscribers only. WhatsApp, Facebook groups, TikTok, Instagram, and SEO all deferred to month 4–6+.

Cinderella Campaign: LinkedIn outreach with visual proof image → referrals from first 2–3 clients. Delivery via client's Klaviyo + monthly PDF report. No platform infrastructure needed.

Returns Audit: Same LinkedIn outreach list as Cinderella. One provocative stat as the entire pitch. Delivery via PDF + 60-minute call. Quarterly renewal converts it to recurring. Naturally feeds Cinderella pipeline.

The product ladder: Returns Audit breaks the ice (one-time, lower commitment) → Cinderella Campaign converts them to monthly retainer (data already shared) → PRI subscription adds intelligence layer on top. One outreach contact can eventually become all three revenue streams.

LinkedIn Targeting — Samurai / SQLite Assignment
Geography · Start Here
Wave 1: New York + California
USD · Highest density
Target USA, not Canada. USD pricing adds 35–40% revenue versus CAD at current rates — at 8 Cinderella clients that is a meaningful difference. The US DTC fashion brand density in the $5–30M revenue range is roughly 10× Canada. Canadian proximity warmth is real but not worth the revenue haircut. Start with New York State and California — these two states concentrate approximately 60% of US DTC fashion brands in the target revenue range. Use the by_State_USA_splits databases. Add Texas and Florida as Wave 2 if volume is insufficient after first outreach run.
Wave 1
New York State · California
Wave 2
Texas · Florida · Illinois
Wave 3
Canada (Toronto, Vancouver) — local warmth angle
Skip entirely
EU / LATAM / MENA until US is validated and capacity exists
Samurai Job 1 · PRI Subscribers
Job 1: Practitioner Reality Index Target
Expected yield: 800–2,000 raw
Ideal person: founder, CEO, or head of marketing at a DTC fashion brand doing $5–30M. Intellectually curious, frustrated with vendor analytics, makes or influences subscription decisions. Not a stylist, journalist, student, or corporate employee at a large group.
Job Title field — include
founder · co-founder · owner · ceo · chief executive · head of marketing · director of marketing · vp marketing · chief marketing officer · vp of marketing · director of ecommerce · head of ecommerce · director of digital marketing
Job Title field — exclude if contains
recruiter · talent · stylist · journalist · editor · student · intern · assistant · coordinator
Summary field — positive signals (score +2 each)
shopify plus · DTC · direct-to-consumer · direct to consumer · fashion brand · apparel brand · clothing brand · ecommerce · e-commerce · klaviyo · 7-figure · 8-figure · million in revenue · omnichannel · wholesale · inventory · email marketing · customer acquisition · retention
Summary field — high-precision bonus signals (score +3 each)
built a fashion brand · launched a brand · scaled our brand · growing our DTC · shopify store · email list · dead stock · markdown · clearance · return rate
Summary field — eliminate if contains
fashion school · studying fashion · fashion student · personal styling · fashion journalist · fashion editor · fashion recruiter · LVMH · PVH · Gap Inc · H&M · Zara · Inditex
Minimum threshold
Job title match + at least 2 summary positive signals
Sample first
Pull 50 records, review manually before full extraction
Export fields
Full name · Job Title · Company Name · Summary (first 300 chars) · LinkedIn URL
Working volume
200 clean contacts sufficient for first outreach wave
Samurai Job 2 · Cinderella + Returns + AI Imagery
Job 2: Service Clients Target
Expected yield: 300–700 raw
Tighter persona than Job 1. Needs someone who directly controls email marketing AND deals with inventory decisions. At brands $5–15M this is usually the founder. At $15–30M it may be a Director of E-commerce or Head of Marketing. Same person can be pitched all three services — Cinderella first, then Returns Audit, then AI imagery.
Job Title field — tighter than Job 1
founder · co-founder · owner · ceo · director of ecommerce · head of ecommerce · director of marketing · head of marketing · email marketing manager · email marketing director
Summary — must contain at least one operational signal
klaviyo · email campaigns · email marketing · segmentation · ecommerce marketing · shopify · inventory · dead stock · clearance · markdown · size run
Summary — and at least one scale signal
shopify plus · 7-figure · 8-figure · million · DTC brand · fashion brand · apparel brand · scaled · growing
Summary — same exclusions as Job 1
fashion school · studying · student · personal styling · journalist · editor · recruiter · LVMH · PVH · Gap Inc · H&M · Zara · Inditex
Minimum threshold
Job title match + 1 operational signal + 1 scale signal
Sample first
Pull 50, review manually — main noise risk is stylists and fashion journalists
Outreach sequence
50 contacts → ~5 replies → ~2 calls → 1 close. Repeat monthly.
Deduplication
Remove records already in Job 1 export. Same person gets one message, not two.
Targeting execution order

Step 1: Run Job 2 first (service clients) — faster revenue, smaller list, easier to manage manually.

Step 2: Run Job 1 in parallel (PRI subscribers) — larger list, longer sales cycle, builds over time.

Step 3: Pull 50-record samples from each before full extraction. Manual review takes 20 minutes and prevents wasting outreach budget on noise.

Step 4: New York first, California second. Run NY outreach for 2 weeks, evaluate reply rates, then add CA. Do not launch both states simultaneously on the first wave — you need signal before scale.

Key rule: Company Name is not a reliable filter for fashion brands — brand names don't contain category keywords. Job Title + Summary is the only combination that works. Do not build a filtered list by company name alone.