Sales Intelligence – Critical calculation, data consistency, competitor matching, AI visibility & German localization issues that should be fixed early

Hi EMR team,

I have now tested Sales Intelligence very extensively across the Self-Service Prospect Report, Full Analysis and Battle Plan, including several different businesses and the German/DACH customer-facing experience.

First: the concept and design are genuinely strong. Sales Intelligence could become one of the most valuable features in EMR for agencies because it connects prospecting, diagnosis and sales in one flow.

However, I have found a number of issues that I believe are important to address while the feature is still relatively new.

Some are clear calculation or data-logic bugs. Others are wording, methodology and localization issues that make modeled or estimated data sound like verified business outcomes.

This matters especially in a white-label environment: the prospect does not know these claims come from EMR. They believe they come from our agency. If a prospect notices one obviously incorrect calculation, trust in the entire report and therefore in the agency can disappear immediately.

I have already sent Support a detailed report and screenshots. I am posting everything here as well because I think these issues are important enough to solve early, before more agencies build their acquisition process around Sales Intelligence.

Environment

• Product: Sales Intelligence

• Areas tested: Self-Service Prospect Report, Full Analysis, Battle Plan

• Customer-facing language: German

• Market: Switzerland / DACH

• White-label agency use case

1. Battle Plan review target calculation is mathematically incorrect

One tested business has:

• Current rating: 4.2★

• Current review count: 398

• Target shown: 4.5★

Battle Plan says:

“2 Bewertungen benötigt” / “2 reviews needed”

It also generates:

“At 4.2 stars, just a few reviews away from 4.5 display threshold.”

This cannot be mathematically correct.

If the true average were exactly 4.20 and every future review were 5 stars:

(4.20 × 398 + 5 × n) / (398 + n) ≥ 4.50

approximately 120 additional 5-star reviews would be required to reach a true 4.50 average.

I understand Google may expose only a rounded rating. If EMR does not know the exact underlying average/rating distribution, then an exact number such as “2 reviews needed” should probably not be shown.

Expected result:
Calculate the requirement from valid underlying data and clearly state the assumption, or show an estimate/range instead.

Example:

“Based on the current review volume, sustained positive review growth is required to move toward 4.5★.”

This is one of the most important bugs because agencies may quote this number directly during a sales conversation.

2. Equal monthly review velocity is incorrectly marked as a deficit

Full Analysis shows:

Prospect: 0 reviews/month
Leader: 0 reviews/month

Yet the prospect receives:

“RÜCKSTAND” / “BEHIND”

and:

“0 frische Bewertungen monatlich mehr benötigt zum Aufholen.”

0 vs. 0 is equal, not behind.

Expected logic:

Prospect < benchmark → behind
Prospect = benchmark → neutral/equal
Prospect > benchmark → ahead

3. Competitor/entity matching appears incorrect

In one Full Analysis for UM bio restaurant, the system selected “Europa” as the market-leading competitor and average Google Maps position #1.

The same competitor card shows:

• 0 reviews

• 0.0★

• 0 reviews/month

• AI visibility 0/3

Yet it is presented as:

Top 3 / average rank #1 / market leader

The analyzed business therefore receives messaging such as:

“4.0 Sterne Vorsprung auf #1.”

This looks like either an upstream data issue, an incorrect entity match, or insufficient competitor validation.

Expected result:
Before an entity becomes the comparison leader, the system should validate that it is an actual relevant local business in the same competitive context.

Generic/non-business entities should not become the main benchmark.

4. Percentage points and percentages are being mixed

One Prospect Report shows approximately:

Leader visibility/traffic capture: 45%
Prospect: 15%

The report then says:

“The top Google Maps position receives 30% more clicks.”

45% − 15% equals 30 percentage points.

Relative to 15%, 45% is 200% higher, not 30% higher.

Expected result:
Distinguish percentage-point differences from relative percentage increases, or avoid converting modeled visibility shares into precise “X% more clicks” claims.

5. Modeled visibility is presented as actual measured traffic

German Prospect Report examples:

“Wettbewerber vor Ihnen erhalten 67% des Traffics, der eigentlich Ihnen gehören sollte.”

Translation:

“Competitors ahead of you receive 67% of the traffic that should actually belong to you.”

Unless Sales Intelligence has access to actual Analytics/Search Console/store traffic for that individual business, this appears to be a ranking-based model, not measured traffic.

Expected wording:

“Based on the measured Google Maps positions, competitors currently have a higher estimated share of local visibility.”

Ideally estimated values should have a tooltip:

“Model-based estimate derived from local ranking positions; not measured website or store traffic.”

6. Modeled visibility is also presented as actual customer loss

Another Prospect Report statement says:

“Ihre Konkurrenten erhalten 67% Ihrer potenziellen Kunden.”

Translation:

“Your competitors receive 67% of your potential customers.”

Ranking visibility does not prove that exactly 67% of potential customers chose competitors.

This is much stronger than the underlying data appears to support.

Better:

“Competitors currently have greater visibility across the analyzed local searches.”

7. Review gap is presented as an objective number of reviews “required”

Example:

Prospect: 126 reviews
Comparison leader: 1,341 reviews

Report displays:

“1215+ weitere benötigt”

Translation:

“1,215+ more required.”

The number is simply the numerical difference to one selected competitor.

It does not necessarily mean the prospect objectively needs exactly 1,215 additional reviews.

Better:

“1,215-review gap to the comparison leader”

or:

“126 vs. 1,341 reviews.”

This keeps the competitive benchmark without turning it into an arbitrary requirement.

8. Local Search Grid red pins are described as actual customers choosing competitors

Current German text:

“Jede rote Stecknadel repräsentiert Kunden, die Wettbewerber statt Sie wählen.”

Translation:

“Each red pin represents customers who choose competitors instead of you.”

A red pin represents weak ranking/visibility at a geographic coordinate. It does not represent an actual customer who made a purchase decision.

Better:

“Each red pin represents a location where the business has limited visibility for the analyzed search.”

9. “Lost revenue” is shown without actual revenue/conversion data

The same Local Search Grid section contains the heading:

“Verlorene Einnahmen” / “Lost revenue.”

Unless actual revenue and conversion data are being used, this is too absolute.

Better:

“Visibility opportunity”

German:

“Ungenutztes Sichtbarkeitspotenzial.”

10. “Potential traffic increase +53%” needs methodology

Full Analysis displays:

“Potenzielle Traffic-Steigerung: +53%.”

It is not clear:

• how this number is calculated,

• whether it is measured or modeled,

• what ranking assumptions are used,

• whether it means website traffic, Maps clicks or estimated visibility.

Better:

“Estimated visibility potential: +53%”

with an info tooltip explaining the calculation.

Methodology transparency would make this feature much more credible.

11. AI visibility conclusions are too absolute

Current German examples:

“Wird nicht von AI empfohlen.”
“Not recommended by AI.”

“Empfiehlt Ihr Unternehmen nicht.”
“Does not recommend your business.”

“Ihr Unternehmen erscheint nicht.”
“Your business does not appear.”

“Sie verlieren potenzielle Kunden an Wettbewerber.”
“You are losing potential customers to competitors.”

“Wettbewerber gewinnen in der AI-Suche.”
“Competitors are winning in AI search.”

Testing selected prompts on ChatGPT, Gemini and Perplexity is valuable.

But not appearing in the sampled queries does not prove that the business is generally never recommended, nor does it prove actual customer loss.

Better:

“The business was not found in the AI queries tested in this analysis.”

For competitors:

“One competitor appeared in at least one of the AI queries tested.”

Ideally the report should expose:

• number of prompts tested

• actual query/category

• location/context

• test date

• platform tested

• notice that AI results can vary between sessions/users

This would make AI Visibility significantly more credible and useful.

12. Battle Plan generates unsupported customer-loss claims

Primary Pitch example:

“That gap is costing you 20–30% of potential customers.”

I cannot see prospect-specific data supporting a precise 20–30% customer-loss estimate.

A rating gap by itself does not prove that a specific business loses exactly that percentage of customers.

Expected result:
Exact customer-loss percentages should only be generated if they are backed by a defined and documented calculation.

Otherwise use qualitative language:

“The rating gap may reduce conversion and competitiveness compared with higher-rated businesses.”

13. Battle Plan creates unsupported 60–90 day result promises

Generated pitch:

“I can help close that gap in 60–90 days with a proven system.”

In the same case the business has 398 reviews and a 4.2 rating.

If the actual goal is 4.5, 60–90 days may be completely unrealistic depending on review velocity.

The system should account for:

• current review count

• current rating

• historical review velocity

• required number of new reviews

• realistic acquisition rate

before generating a timeline.

Better:

“We can build a structured plan to improve review growth and customer engagement over time.”

14. “0.3 stars away from dominating your market” is logically unsupported

Battle Plan hook:

“You’re 0.3 stars away from dominating your market…”

Sales Intelligence itself analyzes multiple dimensions:

• ranking

• review volume

• rating

• review velocity

• response rate

• Local Search Grid

• competitor strength

• sentiment

• AI visibility

Therefore a 0.3-star rating gap alone cannot logically equal “market dominance.”

Better:

“Your online reputation has a solid foundation, with clear opportunities to outperform local competitors.”

15. Prospect Report, Full Analysis and Battle Plan appear to use inconsistent logic

This is probably the most important structural issue.

The three layers should ideally use one shared validated calculation engine for:

• rating calculations

• target ratings

• reviews required

• review velocity

• competitor benchmarks

• ranking gaps

• visibility estimates

• traffic estimates

• AI visibility

• opportunity calculations

The Battle Plan, Cold Call script, Email script, SMS script and Primary Pitch should then consume those validated metrics.

At the moment some sales copy appears to be independently generated in ways that conflict with the actual business data.

16. External statistics are displayed without visible sources

Full Analysis shows claims such as:

93% of customers read reviews before purchasing
5–9% traffic increase for every 0.1-star improvement
52% do not use a business below 4 stars
35% research reviews before visiting a local business

These statistics should ideally include:

Source + publication year + methodology/reference link

through a tooltip or footnote.

In a white-label report this is important because the agency is effectively publishing these claims to its prospects.

17. Sentiment analysis wording overstates small sample sizes

Some themes contain only 7–12 mentions, but the German text says:

“Kunden loben dies durchgängig.”

Translation:

“Customers consistently praise this.”

Better:

“This theme was positively mentioned in 12 analyzed review mentions.”

This is more precise and actually makes the AI sentiment analysis look more professional.

18. Full Analysis intro is unnecessarily alarmist

German intro:

“Eine forensische Analyse Ihres digitalen Fußabdrucks. Wir haben 4 kritische Lücken aufgedeckt, die Sie täglich Kunden kosten.”

Translation:

“A forensic analysis of your digital footprint. We uncovered 4 critical gaps that cost you customers every day.”

Problems:

• “forensic analysis” feels inappropriate for a reputation/marketing report

• “cost you customers every day” is not demonstrated by the data

• it creates unnecessary skepticism before the prospect even sees the evidence

Better:

“A detailed analysis of your digital presence. We identified four areas with clear improvement potential.”

19. German wording contains several unnatural or misleading translations

Examples:

“Marktintelligenz”
Better: “Wettbewerbsanalyse”

“Traffic-Erfassung”
Better: “Geschätzter Sichtbarkeitsanteil”

“Kundenverkehr”
Better depending on context: “lokale Sichtbarkeit” / “Kundengewinnung”

“Tiefe Ergebnisse – Unsichtbar”
This is not natural German.
Better: “Geringe Sichtbarkeit” / “Durchschnittliche Position ~21”

“Das lieben und hassen Ihre Kunden.”
Better: “Was Kunden schätzen und wo Verbesserungspotenzial besteht.”

“Der Kampf”
Better: “Ihr Wettbewerb im lokalen Vergleich.”

“Dominieren Sie die AI-Ära.”
Better: “Verbessern Sie Ihre Präsenz in KI-gestützten Suchergebnissen.”

20. German reports inconsistently mix “AI” and “KI”

Examples use:

• AI-Sichtbarkeit

• AI-Ära

• KI-Assistenten

• AI-Empfehlungswert

For German/DACH reports, terminology should be consistent.

I would recommend KI in German customer-facing copy unless there is a strong branding reason to use AI.

21. Some German sales copy is too aggressive for a professional white-label report

Examples:

“DER KAMPF”

“Im Kampf um Kunden gibt es nur einen Gewinner. Und im Moment sind das nicht Sie.”

Translation:

“There is only one winner in the battle for customers. And right now, it isn’t you.”

“Hören Sie auf, Kunden an Ihre Konkurrenten zu verlieren.”

Translation:

“Stop losing customers to your competitors.”

“Dominieren Sie die AI-Ära.”

This style may work for aggressive direct-response marketing, but it can hurt credibility for agencies positioning themselves as professional consultants.

A more consultative tone would still convert while building more trust.

Example:

“Strengthen your local position and turn identified opportunities into measurable improvements.”

22. Solution section contains overly absolute product/performance claims

Current examples:

“Unsere AI antwortet sofort auf jeden Kunden und steigert Ihr SEO & GEO.”

Translation:

“Our AI instantly responds to every customer and increases your SEO & GEO.”

“Steigen Sie in den Rankings, erfassen Sie den Traffic und sehen Sie Ihr Geschäft wachsen.”

Translation:

“Rise in the rankings, capture the traffic and watch your business grow.”

These imply guaranteed outcomes.

Better:

“Use AI-assisted responses to manage reviews faster and more consistently.”

“Systematically improve reviews, customer interaction and local visibility.”

23. “Market leader” wording is sometimes too broad

Example:

A prospect has 4.3★ versus a comparison competitor with 4.0★.

The report says:

“Sie sind der Marktführer!” / “You are the market leader!”

Having the highest star rating among the selected comparison set does not automatically mean the business is the overall market leader.

Better:

“Highest Google rating in this comparison.”

24. White-label agencies need more control over report copy

The new customizable Trust Score status messages are a very good step.

I strongly recommend extending this concept to all major Sales Intelligence customer-facing copy, per language.

Agencies should ideally be able to customize:

• report headlines

• traffic/visibility terminology

• competitor messaging

• review-gap labels

• AI visibility wording

• Full Analysis headings

• solution descriptions

• Battle Plan tone

• disclaimers/methodology text

This would allow agencies to choose between:

aggressive sales / consultative / compliance-focused

while keeping the actual analytics engine unchanged.

25. Strong conversion opportunity: connect every detected issue directly to an EMR solution

I think this could materially improve conversion.

Instead of relying on fear-based claims, the report could dynamically create a:

“How to improve this” / “How we can solve this”

section.

Examples:

0% response rate
→ AI-assisted review responses

Low review velocity
→ automated review requests via email/SMS/QR/NFC

Weak local rankings
→ Local Search Grid monitoring and local reputation workflow

Negative sentiment themes
→ sentiment monitoring/reputation management

Limited AI visibility
→ AI visibility monitoring

This creates a much stronger sales journey:

Data → Diagnosis → Problem → Recommended action → EMR solution → CTA

The prospect then understands exactly why they would need the platform.

Why I think these issues should be addressed early

Sales Intelligence is still relatively new, which is exactly why I think this is the right moment to solve these problems.

The feature has the potential to become a major differentiator for EMR.

But its value depends heavily on trust in the data.

When an agency sends a report under its own brand, even one obvious issue such as:

4.2★ + 398 reviews → only 2 reviews needed for 4.5★

can cause the prospect to question every other figure in the report.

The same applies when modeled visibility is presented as actual traffic, ranking points are described as real lost customers, or generic entities appear as market-leading competitors.

These are not small cosmetic issues when Sales Intelligence is used as a sales tool. They directly affect:

agency credibility, prospect trust, conversion and long-term adoption of the feature.

I would therefore strongly recommend prioritizing the calculation/data-consistency issues first, followed by methodology transparency and localization/copy improvements.

I have already provided Support with screenshots and a detailed PDF covering the examples above, and I am happy to provide exact report IDs or additional test cases privately to the product/engineering team.

I’m raising this because I think Sales Intelligence can become an exceptionally strong part of EMR if these foundations are made reliable now, while the product is still evolving.